[CostModel][X86] Add CostKinds handling for mul ops

This was achieved using the 'cost-tables vs llvm-mca' script D103695

Also fix a missing pmullw v16i16 half-rate throughput as znver1 double-pumps - matches numbers from AMD SoG + Agner

diff --git a/llvm/lib/Target/X86/X86ScheduleZnver1.td b/llvm/lib/Target/X86/X86ScheduleZnver1.td
index 463ec90..8003bba 100644 (file)
--- a/llvm/lib/Target/X86/X86ScheduleZnver1.td
+++ b/llvm/lib/Target/X86/X86ScheduleZnver1.td
@@ -422,7 +422,7 @@ defm : ZnWriteResFpuPair<WriteVecALUY,    [ZnFPU],   1, [2], 2>;
 defm : X86WriteResPairUnsupported<WriteVecALUZ>;
 defm : ZnWriteResFpuPair<WriteVecIMul,    [ZnFPU0],  4>;
 defm : ZnWriteResFpuPair<WriteVecIMulX,   [ZnFPU0],  4>;
-defm : ZnWriteResFpuPair<WriteVecIMulY,   [ZnFPU0],  4>;
+defm : ZnWriteResFpuPair<WriteVecIMulY,   [ZnFPU0],  4, [2], 2>;
 defm : X86WriteResPairUnsupported<WriteVecIMulZ>;
 defm : ZnWriteResFpuPair<WritePMULLD,     [ZnFPU0],  4, [2]>;
 defm : ZnWriteResFpuPair<WritePMULLDY,    [ZnFPU0],  4, [4], 2>;

diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index a3edca0..5126796 100644 (file)
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -341,8 +341,8 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
         return LT.first * KindCost.value();
 
   static const CostKindTblEntry SLMCostTable[] = {
-    { ISD::MUL,   MVT::v4i32, { 11 } }, // pmulld
-    { ISD::MUL,   MVT::v8i16, { 2  } }, // pmullw
+    { ISD::MUL,   MVT::v4i32, { 11, 11, 1, 7 } }, // pmulld
+    { ISD::MUL,   MVT::v8i16, {  2,  5, 1, 1 } }, // pmullw
     { ISD::FMUL,  MVT::f64,   {  2,  5, 1, 1 } }, // mulsd
     { ISD::FMUL,  MVT::f32,   {  1,  4, 1, 1 } }, // mulss
     { ISD::FMUL,  MVT::v2f64, {  4,  7, 1, 1 } }, // mulpd
@@ -358,7 +358,7 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
     // slm muldq version throughput is 2 and addq throughput 4
     // thus: 3X2 (muldq throughput) + 3X1 (shift throughput) +
     //       3X4 (addq throughput) = 17
-    { ISD::MUL,   MVT::v2i64, { 17 } },
+    { ISD::MUL,   MVT::v2i64, { 17, 22, 9, 9 } },
     // slm addq\subq throughput is 4
     { ISD::ADD,   MVT::v2i64, {  4,  2, 1, 2 } },
     { ISD::SUB,   MVT::v2i64, {  4,  2, 1, 2 } },
@@ -629,9 +629,9 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
         return LT.first * KindCost.value();
 
   static const CostKindTblEntry AVX512DQCostTable[] = {
-    { ISD::MUL,  MVT::v2i64, { 2 } }, // pmullq
-    { ISD::MUL,  MVT::v4i64, { 2 } }, // pmullq
-    { ISD::MUL,  MVT::v8i64, { 2 } }  // pmullq
+    { ISD::MUL,  MVT::v2i64, { 2, 15, 1, 3 } }, // pmullq
+    { ISD::MUL,  MVT::v4i64, { 2, 15, 1, 3 } }, // pmullq
+    { ISD::MUL,  MVT::v8i64, { 2, 15, 1, 3 } }  // pmullq
   };
 
   // Look for AVX512DQ lowering tricks for custom cases.
@@ -656,6 +656,8 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
     { ISD::SUB,   MVT::v64i8,   {  1,  1, 1, 1 } }, // psubb
     { ISD::SUB,   MVT::v32i16,  {  1,  1, 1, 1 } }, // psubw
 
+    { ISD::MUL,   MVT::v32i16,  {  1,  5, 1, 1 } }, // pmullw
+
     { ISD::SUB,   MVT::v32i8,   {  1,  1, 1, 1 } }, // psubb
     { ISD::SUB,   MVT::v16i16,  {  1,  1, 1, 1 } }, // psubw
     { ISD::SUB,   MVT::v8i32,   {  1,  1, 1, 1 } }, // psubd
@@ -711,10 +713,10 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
     { ISD::XOR,     MVT::v8i32,   {  1,  1, 1, 1 } },
     { ISD::XOR,     MVT::v4i64,   {  1,  1, 1, 1 } },
 
-    { ISD::MUL,     MVT::v16i32,  {  1 } }, // pmulld (Skylake from agner.org)
-    { ISD::MUL,     MVT::v8i32,   {  1 } }, // pmulld (Skylake from agner.org)
-    { ISD::MUL,     MVT::v4i32,   {  1 } }, // pmulld (Skylake from agner.org)
-    { ISD::MUL,     MVT::v8i64,   {  6 } }, // 3*pmuludq/3*shift/2*add
+    { ISD::MUL,     MVT::v16i32,  {  1, 10, 1, 2 } }, // pmulld (Skylake from agner.org)
+    { ISD::MUL,     MVT::v8i32,   {  1, 10, 1, 2 } }, // pmulld (Skylake from agner.org)
+    { ISD::MUL,     MVT::v4i32,   {  1, 10, 1, 2 } }, // pmulld (Skylake from agner.org)
+    { ISD::MUL,     MVT::v8i64,   {  6,  9, 8, 8 } }, // 3*pmuludq/3*shift/2*add
     { ISD::MUL,     MVT::i64,     {  1 } }, // Skylake from http://www.agner.org/
 
     { ISD::FNEG,    MVT::v8f64,   {  1,  1, 1, 2 } }, // Skylake from http://www.agner.org/
@@ -906,9 +908,11 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
     { ISD::SUB,  MVT::v4i64,   {  1,  1, 1, 2 } }, // psubq
     { ISD::ADD,  MVT::v4i64,   {  1,  1, 1, 2 } }, // paddq
 
-    { ISD::MUL,  MVT::v16i16,  {  1 } }, // pmullw
-    { ISD::MUL,  MVT::v8i32,   {  4 } }, // pmulld
-    { ISD::MUL,  MVT::v4i64,   {  6 } }, // 3*pmuludq/3*shift/2*add
+    { ISD::MUL,  MVT::v16i16,  {  2,  5, 1, 1 } }, // pmullw
+    { ISD::MUL,  MVT::v8i32,   {  4, 10, 1, 2 } }, // pmulld
+    { ISD::MUL,  MVT::v4i32,   {  2, 10, 1, 2 } }, // pmulld
+    { ISD::MUL,  MVT::v4i64,   {  6, 10, 8,13 } }, // 3*pmuludq/3*shift/2*add
+    { ISD::MUL,  MVT::v2i64,   {  6, 10, 8, 8 } }, // 3*pmuludq/3*shift/2*add
 
     { ISD::FNEG, MVT::v4f64,   {  1,  1, 1, 2 } }, // vxorpd
     { ISD::FNEG, MVT::v8f32,   {  1,  1, 1, 2 } }, // vxorps
@@ -952,9 +956,10 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
     // We don't have to scalarize unsupported ops. We can issue two half-sized
     // operations and we only need to extract the upper YMM half.
     // Two ops + 1 extract + 1 insert = 4.
-    { ISD::MUL,     MVT::v16i16,  {  4 } },
-    { ISD::MUL,     MVT::v8i32,   {  5 } }, // BTVER2 from http://www.agner.org/
-    { ISD::MUL,     MVT::v4i64,   { 12 } },
+    { ISD::MUL,     MVT::v16i16,  {  4,  8,  5,  6 } }, // pmullw + split
+    { ISD::MUL,     MVT::v8i32,   {  5,  8,  5, 10 } }, // pmulld + split
+    { ISD::MUL,     MVT::v4i32,   {  2,  5,  1,  3 } }, // pmulld
+    { ISD::MUL,     MVT::v4i64,   { 12, 15, 19, 20 } },
 
     { ISD::AND,     MVT::v32i8,   {  1,  1, 1, 2 } }, // vandps
     { ISD::AND,     MVT::v16i16,  {  1,  1, 1, 2 } }, // vandps
@@ -1062,7 +1067,7 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
     { ISD::FDIV,  MVT::f64,   { 22, 22, 1, 1 } }, // Nehalem from http://www.agner.org/
     { ISD::FDIV,  MVT::v2f64, { 22, 22, 1, 1 } }, // Nehalem from http://www.agner.org/
 
-    { ISD::MUL,   MVT::v2i64, {  6 } }  // 3*pmuludq/3*shift/2*add
+    { ISD::MUL,   MVT::v2i64, {  6, 10,10,10 } }  // 3*pmuludq/3*shift/2*add
   };
 
   if (ST->hasSSE42())
@@ -1083,7 +1088,7 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
     { ISD::SRA,  MVT::v16i8,  { 21 } }, // pblendvb sequence.
     { ISD::SRA,  MVT::v8i16,  { 13 } }, // pblendvb sequence.
 
-    { ISD::MUL,  MVT::v4i32,  {  2 } }  // pmulld (Nehalem from agner.org)
+    { ISD::MUL,  MVT::v4i32,  {  2, 11, 1, 1 } }  // pmulld (Nehalem from agner.org)
   };
 
   if (ST->hasSSE41())
@@ -1127,9 +1132,9 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
     { ISD::ADD,  MVT::v2i64,  {  1,  2, 1, 2 } }, // paddq
     { ISD::SUB,  MVT::v2i64,  {  1,  2, 1, 2 } }, // psubq
 
-    { ISD::MUL,  MVT::v8i16,  {  1 } }, // pmullw
-    { ISD::MUL,  MVT::v4i32,  {  6 } }, // 3*pmuludq/4*shuffle
-    { ISD::MUL,  MVT::v2i64,  {  8 } }, // 3*pmuludq/3*shift/2*add
+    { ISD::MUL,  MVT::v8i16,  {  1,  5, 1, 1 } }, // pmullw
+    { ISD::MUL,  MVT::v4i32,  {  6,  8, 7, 7 } }, // 3*pmuludq/4*shuffle
+    { ISD::MUL,  MVT::v2i64,  {  8, 10, 8, 8 } }, // 3*pmuludq/3*shift/2*add
 
     { ISD::FDIV, MVT::f32,    { 23, 23, 1, 1 } }, // Pentium IV from http://www.agner.org/
     { ISD::FDIV, MVT::v4f32,  { 39, 39, 1, 1 } }, // Pentium IV from http://www.agner.org/

diff --git a/llvm/test/Analysis/CostModel/X86/arith-fix.ll b/llvm/test/Analysis/CostModel/X86/arith-fix.ll
index 6610b81..1febfee 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/arith-fix.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-fix.ll
@@ -103,9 +103,9 @@ define i32 @smul(i32 %arg) {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V16I16 = call <16 x i16> @llvm.smul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V32I16 = call <32 x i16> @llvm.smul.fix.v32i16(<32 x i16> undef, <32 x i16> undef, i32 3)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = call i8 @llvm.smul.fix.i8(i8 undef, i8 undef, i32 3)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I8 = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32I8 = call <32 x i8> @llvm.smul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V64I8 = call <64 x i8> @llvm.smul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V32I8 = call <32 x i8> @llvm.smul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 58 for instruction: %V64I8 = call <64 x i8> @llvm.smul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'smul'
@@ -122,7 +122,7 @@ define i32 @smul(i32 %arg) {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.smul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V32I16 = call <32 x i16> @llvm.smul.fix.v32i16(<32 x i16> undef, <32 x i16> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = call i8 @llvm.smul.fix.i8(i8 undef, i8 undef, i32 3)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I8 = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V32I8 = call <32 x i8> @llvm.smul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V64I8 = call <64 x i8> @llvm.smul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -141,7 +141,7 @@ define i32 @smul(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.smul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V32I16 = call <32 x i16> @llvm.smul.fix.v32i16(<32 x i16> undef, <32 x i16> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = call i8 @llvm.smul.fix.i8(i8 undef, i8 undef, i32 3)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I8 = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I8 = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32I8 = call <32 x i8> @llvm.smul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 45 for instruction: %V64I8 = call <64 x i8> @llvm.smul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -160,7 +160,7 @@ define i32 @smul(i32 %arg) {
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.smul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V32I16 = call <32 x i16> @llvm.smul.fix.v32i16(<32 x i16> undef, <32 x i16> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = call i8 @llvm.smul.fix.i8(i8 undef, i8 undef, i32 3)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I8 = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V32I8 = call <32 x i8> @llvm.smul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V64I8 = call <64 x i8> @llvm.smul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -337,9 +337,9 @@ define i32 @umul(i32 %arg) {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V16I16 = call <16 x i16> @llvm.umul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V32I16 = call <32 x i16> @llvm.umul.fix.v32i16(<32 x i16> undef, <32 x i16> undef, i32 3)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = call i8 @llvm.umul.fix.i8(i8 undef, i8 undef, i32 3)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I8 = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32I8 = call <32 x i8> @llvm.umul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V64I8 = call <64 x i8> @llvm.umul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V32I8 = call <32 x i8> @llvm.umul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 58 for instruction: %V64I8 = call <64 x i8> @llvm.umul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'umul'
@@ -356,7 +356,7 @@ define i32 @umul(i32 %arg) {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.umul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V32I16 = call <32 x i16> @llvm.umul.fix.v32i16(<32 x i16> undef, <32 x i16> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = call i8 @llvm.umul.fix.i8(i8 undef, i8 undef, i32 3)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I8 = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V32I8 = call <32 x i8> @llvm.umul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V64I8 = call <64 x i8> @llvm.umul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -375,7 +375,7 @@ define i32 @umul(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.umul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V32I16 = call <32 x i16> @llvm.umul.fix.v32i16(<32 x i16> undef, <32 x i16> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = call i8 @llvm.umul.fix.i8(i8 undef, i8 undef, i32 3)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I8 = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I8 = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32I8 = call <32 x i8> @llvm.umul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 45 for instruction: %V64I8 = call <64 x i8> @llvm.umul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -394,7 +394,7 @@ define i32 @umul(i32 %arg) {
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.umul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V32I16 = call <32 x i16> @llvm.umul.fix.v32i16(<32 x i16> undef, <32 x i16> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = call i8 @llvm.umul.fix.i8(i8 undef, i8 undef, i32 3)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I8 = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> undef, <16 x i8> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V32I8 = call <32 x i8> @llvm.umul.fix.v32i8(<32 x i8> undef, <32 x i8> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V64I8 = call <64 x i8> @llvm.umul.fix.v64i8(<64 x i8> undef, <64 x i8> undef, i32 3)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef

diff --git a/llvm/test/Analysis/CostModel/X86/arith-int-codesize.ll b/llvm/test/Analysis/CostModel/X86/arith-int-codesize.ll
index ba0dc69..2dbdd06 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/arith-int-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-int-codesize.ll
@@ -1,15 +1,15 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX512F
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX512BW
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=CHECK,AVX512F
-;
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mcpu=btver2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE,SSE2
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE,SSE42
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=AVX512DQ
+;
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mcpu=slm | FileCheck %s --check-prefixes=SSE,SLM
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mcpu=goldmont | FileCheck %s --check-prefixes=SSE,SSE42
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mtriple=x86_64-- -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
 
 define i32 @add(i32 %arg) {
 ; SSE-LABEL: 'add'
@@ -107,6 +107,25 @@ define i32 @add(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = add <64 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'add'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = add i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = add <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = add <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = add <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = add i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = add <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = add <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = add <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = add i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = add <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = add <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32I16 = add <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = add i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = add <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I8 = add <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64I8 = add <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = add i64 undef, undef
   %V2I64 = add <2 x i64> undef, undef
   %V4I64 = add <4 x i64> undef, undef
@@ -226,6 +245,25 @@ define i32 @sub(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = sub <64 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'sub'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sub i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = sub <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = sub <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = sub <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sub i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = sub <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = sub <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = sub <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sub i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = sub <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = sub <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32I16 = sub <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = sub i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = sub <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I8 = sub <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64I8 = sub <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = sub i64 undef, undef
   %V2I64 = sub <2 x i64> undef, undef
   %V4I64 = sub <4 x i64> undef, undef
@@ -354,6 +392,32 @@ define i32 @or(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I1 = or <64 x i1> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'or'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = or i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = or <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = or <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = or <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = or i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = or <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = or <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = or <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = or i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = or <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = or <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = or <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = or i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = or <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I8 = or <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = or <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I1 = or i1 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I1 = or <2 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I1 = or <4 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I1 = or <8 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I1 = or <16 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I1 = or <32 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64I1 = or <64 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = or i64 undef, undef
   %V2I64 = or <2 x i64> undef, undef
   %V4I64 = or <4 x i64> undef, undef
@@ -490,6 +554,32 @@ define i32 @xor(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I1 = xor <64 x i1> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'xor'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = xor i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = xor <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = xor <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = xor <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = xor i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = xor <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = xor <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = xor <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = xor i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = xor <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = xor <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = xor <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = xor i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = xor <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I8 = xor <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = xor <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I1 = xor i1 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I1 = xor <2 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I1 = xor <4 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I1 = xor <8 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I1 = xor <16 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I1 = xor <32 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64I1 = xor <64 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = xor i64 undef, undef
   %V2I64 = xor <2 x i64> undef, undef
   %V4I64 = xor <4 x i64> undef, undef
@@ -626,6 +716,32 @@ define i32 @and(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I1 = and <64 x i1> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'and'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = and i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = and <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = and <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = and <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = and i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = and <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = and <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = and <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = and i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = and <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = and <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = and <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = and i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = and <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I8 = and <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = and <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I1 = and i1 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I1 = and <2 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I1 = and <4 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I1 = and <8 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I1 = and <16 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I1 = and <32 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64I1 = and <64 x i1> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = and i64 undef, undef
   %V2I64 = and <2 x i64> undef, undef
   %V4I64 = and <4 x i64> undef, undef
@@ -658,27 +774,181 @@ define i32 @and(i32 %arg) {
 }
 
 define i32 @mul(i32 %arg) {
-; CHECK-LABEL: 'mul'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = mul <2 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = mul <4 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = mul <8 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = mul <8 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'mul'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'mul'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'mul'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'mul'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512F-LABEL: 'mul'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512BW-LABEL: 'mul'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512DQ-LABEL: 'mul'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SLM-LABEL: 'mul'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %I64 = mul i64 undef, undef
   %V2I64 = mul <2 x i64> undef, undef
@@ -710,9 +980,33 @@ define i32 @mul(i32 %arg) {
 ; A <2 x i64> vector multiply is implemented using
 ; 3 PMULUDQ and 2 PADDS and 4 shifts.
 define void @mul_2i32() {
-; CHECK-LABEL: 'mul_2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'mul_2i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %A0 = mul <2 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'mul_2i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX-LABEL: 'mul_2i32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX512F-LABEL: 'mul_2i32'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX512BW-LABEL: 'mul_2i32'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX512DQ-LABEL: 'mul_2i32'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SLM-LABEL: 'mul_2i32'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %A0 = mul <2 x i32> undef, undef
 

diff --git a/llvm/test/Analysis/CostModel/X86/arith-int-latency.ll b/llvm/test/Analysis/CostModel/X86/arith-int-latency.ll
index 7e5a4b3..1a13174 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/arith-int-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-int-latency.ll
@@ -1,15 +1,15 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512F
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512BW
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512F
-;
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mcpu=btver2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE,SSE2
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE,SSE42
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=AVX512,AVX512DQ
+;
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mcpu=slm | FileCheck %s --check-prefixes=SSE,SLM
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mcpu=goldmont | FileCheck %s --check-prefixes=SSE,SSE42
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mtriple=x86_64-- -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
 
 define i32 @add(i32 %arg) {
 ; SSE-LABEL: 'add'
@@ -107,6 +107,25 @@ define i32 @add(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = add <64 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'add'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = add i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = add <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = add <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = add <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = add i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = add <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = add <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = add <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = add i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = add <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = add <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32I16 = add <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = add i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = add <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I8 = add <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64I8 = add <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = add i64 undef, undef
   %V2I64 = add <2 x i64> undef, undef
   %V4I64 = add <4 x i64> undef, undef
@@ -226,6 +245,25 @@ define i32 @sub(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = sub <64 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'sub'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sub i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = sub <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = sub <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = sub <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sub i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = sub <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = sub <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = sub <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sub i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = sub <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = sub <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32I16 = sub <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = sub i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = sub <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I8 = sub <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64I8 = sub <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = sub i64 undef, undef
   %V2I64 = sub <2 x i64> undef, undef
   %V4I64 = sub <4 x i64> undef, undef
@@ -580,27 +618,181 @@ define i32 @and(i32 %arg) {
 }
 
 define i32 @mul(i32 %arg) {
-; CHECK-LABEL: 'mul'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = mul <2 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = mul <4 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = mul <8 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = mul <8 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'mul'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'mul'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 44 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'mul'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 34 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'mul'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512F-LABEL: 'mul'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512BW-LABEL: 'mul'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512DQ-LABEL: 'mul'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SLM-LABEL: 'mul'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 44 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 88 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 44 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %I64 = mul i64 undef, undef
   %V2I64 = mul <2 x i64> undef, undef
@@ -632,9 +824,29 @@ define i32 @mul(i32 %arg) {
 ; A <2 x i64> vector multiply is implemented using
 ; 3 PMULUDQ and 2 PADDS and 4 shifts.
 define void @mul_2i32() {
-; CHECK-LABEL: 'mul_2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'mul_2i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %A0 = mul <2 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'mul_2i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %A0 = mul <2 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX1-LABEL: 'mul_2i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX2-LABEL: 'mul_2i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX512-LABEL: 'mul_2i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SLM-LABEL: 'mul_2i32'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %A0 = mul <2 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %A0 = mul <2 x i32> undef, undef
 

diff --git a/llvm/test/Analysis/CostModel/X86/arith-int-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/arith-int-sizelatency.ll
index eef9156..b7732bb 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/arith-int-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-int-sizelatency.ll
@@ -1,15 +1,15 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512F
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512BW
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512F
-;
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mcpu=btver2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE,SSE2
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE,SSE42
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=AVX512,AVX512DQ
+;
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mcpu=slm | FileCheck %s --check-prefixes=SSE,SLM
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mcpu=goldmont | FileCheck %s --check-prefixes=SSE,SSE42
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mtriple=x86_64-- -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
 
 define i32 @add(i32 %arg) {
 ; SSE-LABEL: 'add'
@@ -107,6 +107,25 @@ define i32 @add(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = add <64 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'add'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = add i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = add <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4I64 = add <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = add <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = add i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = add <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = add <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = add <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = add i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = add <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = add <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32I16 = add <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = add i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = add <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I8 = add <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64I8 = add <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = add i64 undef, undef
   %V2I64 = add <2 x i64> undef, undef
   %V4I64 = add <4 x i64> undef, undef
@@ -226,6 +245,25 @@ define i32 @sub(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64I8 = sub <64 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; AVX512DQ-LABEL: 'sub'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sub i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = sub <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4I64 = sub <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = sub <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sub i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = sub <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = sub <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = sub <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sub i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = sub <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = sub <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32I16 = sub <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = sub i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = sub <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I8 = sub <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64I8 = sub <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = sub i64 undef, undef
   %V2I64 = sub <2 x i64> undef, undef
   %V4I64 = sub <4 x i64> undef, undef
@@ -580,27 +618,181 @@ define i32 @and(i32 %arg) {
 }
 
 define i32 @mul(i32 %arg) {
-; CHECK-LABEL: 'mul'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2I64 = mul <2 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I64 = mul <4 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I64 = mul <8 x i64> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I32 = mul <8 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'mul'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'mul'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'mul'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'mul'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512F-LABEL: 'mul'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512BW-LABEL: 'mul'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512DQ-LABEL: 'mul'
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SLM-LABEL: 'mul'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V2I64 = mul <2 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4I32 = mul <4 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8I32 = mul <8 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V16I32 = mul <16 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %I64 = mul i64 undef, undef
   %V2I64 = mul <2 x i64> undef, undef
@@ -632,9 +824,29 @@ define i32 @mul(i32 %arg) {
 ; A <2 x i64> vector multiply is implemented using
 ; 3 PMULUDQ and 2 PADDS and 4 shifts.
 define void @mul_2i32() {
-; CHECK-LABEL: 'mul_2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'mul_2i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %A0 = mul <2 x i32> undef, undef
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'mul_2i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %A0 = mul <2 x i32> undef, undef
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX1-LABEL: 'mul_2i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX2-LABEL: 'mul_2i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX512-LABEL: 'mul_2i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %A0 = mul <2 x i32> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SLM-LABEL: 'mul_2i32'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %A0 = mul <2 x i32> undef, undef
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %A0 = mul <2 x i32> undef, undef
 

diff --git a/llvm/test/Analysis/CostModel/X86/arith-int.ll b/llvm/test/Analysis/CostModel/X86/arith-int.ll
index 58196b0..c729d88 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/arith-int.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-int.ll
@@ -1003,15 +1003,15 @@ define i32 @mul(i32 %arg) {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16I32 = mul <16 x i32> undef, undef
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I16 = mul <32 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = mul <32 x i16> undef, undef
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I8 = mul <16 x i8> undef, undef
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32I8 = mul <32 x i8> undef, undef
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V64I8 = mul <64 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32I8 = mul <32 x i8> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V64I8 = mul <64 x i8> undef, undef
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'mul'
@@ -1025,13 +1025,13 @@ define i32 @mul(i32 %arg) {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I16 = mul <32 x i16> undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16I8 = mul <16 x i8> undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32I8 = mul <32 x i8> undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V64I8 = mul <64 x i8> undef, undef
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1047,13 +1047,13 @@ define i32 @mul(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32I16 = mul <32 x i16> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = mul <2 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4I8 = mul <4 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8I8 = mul <8 x i8> undef, undef
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I8 = mul <16 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32I8 = mul <32 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V64I8 = mul <64 x i8> undef, undef
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1069,13 +1069,13 @@ define i32 @mul(i32 %arg) {
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I32 = mul <16 x i32> undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = mul <8 x i16> undef, undef
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16I16 = mul <16 x i16> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = mul <16 x i16> undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32I16 = mul <32 x i16> undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = mul <2 x i8> undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = mul <4 x i8> undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8I8 = mul <8 x i8> undef, undef
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16I8 = mul <16 x i8> undef, undef
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16I8 = mul <16 x i8> undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32I8 = mul <32 x i8> undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V64I8 = mul <64 x i8> undef, undef
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef

diff --git a/llvm/test/Analysis/CostModel/X86/arith-overflow.ll b/llvm/test/Analysis/CostModel/X86/arith-overflow.ll
index eeeac17..8ec7d48 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/arith-overflow.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-overflow.ll
@@ -1059,9 +1059,9 @@ define i32 @smul(i32 %arg) {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.smul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V32I16 = call { <32 x i16>, <32 x i1> } @llvm.smul.with.overflow.v32i16(<32 x i16> undef, <32 x i16> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %I8 = call { i8, i1 } @llvm.smul.with.overflow.i8(i8 undef, i8 undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.smul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.smul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 60 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.smul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.smul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.smul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.smul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'smul'
@@ -1078,7 +1078,7 @@ define i32 @smul(i32 %arg) {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.smul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V32I16 = call { <32 x i16>, <32 x i1> } @llvm.smul.with.overflow.v32i16(<32 x i16> undef, <32 x i16> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %I8 = call { i8, i1 } @llvm.smul.with.overflow.i8(i8 undef, i8 undef)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.smul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.smul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.smul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 65 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.smul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1097,7 +1097,7 @@ define i32 @smul(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.smul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = call { <32 x i16>, <32 x i1> } @llvm.smul.with.overflow.v32i16(<32 x i16> undef, <32 x i16> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %I8 = call { i8, i1 } @llvm.smul.with.overflow.i8(i8 undef, i8 undef)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.smul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.smul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.smul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.smul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1116,7 +1116,7 @@ define i32 @smul(i32 %arg) {
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.smul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V32I16 = call { <32 x i16>, <32 x i1> } @llvm.smul.with.overflow.v32i16(<32 x i16> undef, <32 x i16> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %I8 = call { i8, i1 } @llvm.smul.with.overflow.i8(i8 undef, i8 undef)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.smul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.smul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.smul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 65 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.smul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1297,9 +1297,9 @@ define i32 @umul(i32 %arg) {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.umul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 52 for instruction: %V32I16 = call { <32 x i16>, <32 x i1> } @llvm.umul.with.overflow.v32i16(<32 x i16> undef, <32 x i16> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %I8 = call { i8, i1 } @llvm.umul.with.overflow.i8(i8 undef, i8 undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.umul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.umul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 52 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.umul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.umul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.umul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.umul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'umul'
@@ -1316,7 +1316,7 @@ define i32 @umul(i32 %arg) {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.umul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V32I16 = call { <32 x i16>, <32 x i1> } @llvm.umul.with.overflow.v32i16(<32 x i16> undef, <32 x i16> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %I8 = call { i8, i1 } @llvm.umul.with.overflow.i8(i8 undef, i8 undef)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.umul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.umul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.umul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 57 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.umul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1335,7 +1335,7 @@ define i32 @umul(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.umul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V32I16 = call { <32 x i16>, <32 x i1> } @llvm.umul.with.overflow.v32i16(<32 x i16> undef, <32 x i16> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %I8 = call { i8, i1 } @llvm.umul.with.overflow.i8(i8 undef, i8 undef)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.umul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.umul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.umul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 43 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.umul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1354,7 +1354,7 @@ define i32 @umul(i32 %arg) {
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.umul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V32I16 = call { <32 x i16>, <32 x i1> } @llvm.umul.with.overflow.v32i16(<32 x i16> undef, <32 x i16> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %I8 = call { i8, i1 } @llvm.umul.with.overflow.i8(i8 undef, i8 undef)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.umul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I8 = call { <16 x i8>, <16 x i1> } @llvm.umul.with.overflow.v16i8(<16 x i8> undef, <16 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32I8 = call { <32 x i8>, <32 x i1> } @llvm.umul.with.overflow.v32i8(<32 x i8> undef, <32 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 57 for instruction: %V64I8 = call { <64 x i8>, <64 x i1> } @llvm.umul.with.overflow.v64i8(<64 x i8> undef, <64 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef

diff --git a/llvm/test/Analysis/CostModel/X86/fshl.ll b/llvm/test/Analysis/CostModel/X86/fshl.ll
index 7ec788d..c79583a 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/fshl.ll
+++ b/llvm/test/Analysis/CostModel/X86/fshl.ll
@@ -803,14 +803,14 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
 ; AVX2-LABEL: 'constant_funnel_i16'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 7)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'constant_funnel_i16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 7)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -824,7 +824,7 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
 ; AVX512DQ-LABEL: 'constant_funnel_i16'
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 7)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -1966,14 +1966,14 @@ define void @constant_rotate_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
 ; AVX2-LABEL: 'constant_rotate_i16'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %a16, i16 7)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %a128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %a512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %a512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'constant_rotate_i16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %a16, i16 7)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %a128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %a512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -1987,7 +1987,7 @@ define void @constant_rotate_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
 ; AVX512DQ-LABEL: 'constant_rotate_i16'
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %a16, i16 7)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %a128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %a512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;

diff --git a/llvm/test/Analysis/CostModel/X86/fshr.ll b/llvm/test/Analysis/CostModel/X86/fshr.ll
index 80ab1bc..9b73dbc 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/fshr.ll
+++ b/llvm/test/Analysis/CostModel/X86/fshr.ll
@@ -803,14 +803,14 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
 ; AVX2-LABEL: 'constant_funnel_i16'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 7)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'constant_funnel_i16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 7)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -824,7 +824,7 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
 ; AVX512DQ-LABEL: 'constant_funnel_i16'
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 7)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -1967,14 +1967,14 @@ define void @constant_rotate_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
 ; AVX2-LABEL: 'constant_rotate_i16'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %a16, i16 7)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %a128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %a512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %a512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'constant_rotate_i16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %a16, i16 7)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %a128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %a512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -1988,7 +1988,7 @@ define void @constant_rotate_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
 ; AVX512DQ-LABEL: 'constant_rotate_i16'
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %a16, i16 7)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %a128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %a256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %a512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;

diff --git a/llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll b/llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll
index 01b1b7f..ac61f88 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll
+++ b/llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll
@@ -56,17 +56,17 @@ define void @umul(i32 %a, i32 %b, <16 x i32> %va, <16 x i32> %vb) {
 ;
 ; LATE-LABEL: 'umul'
 ; LATE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %s = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
-; LATE-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %v = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> %va, <16 x i32> %vb)
+; LATE-NEXT:  Cost Model: Found an estimated cost of 86 for instruction: %v = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> %va, <16 x i32> %vb)
 ; LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SIZE-LABEL: 'umul'
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %s = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %v = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> %va, <16 x i32> %vb)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 70 for instruction: %v = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> %va, <16 x i32> %vb)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SIZE_LATE-LABEL: 'umul'
 ; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %s = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
-; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %v = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> %va, <16 x i32> %vb)
+; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 70 for instruction: %v = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> %va, <16 x i32> %vb)
 ; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %s = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)

diff --git a/llvm/test/Analysis/CostModel/X86/mul.ll b/llvm/test/Analysis/CostModel/X86/mul.ll
index 63329e5..3cfccbb 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/mul.ll
+++ b/llvm/test/Analysis/CostModel/X86/mul.ll
@@ -99,12 +99,12 @@ define i32 @mul_constpow2() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16i32 = mul <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = mul <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i16 = mul <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i16 = mul <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = mul <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32i16 = mul <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i8 = mul <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i8 = mul <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V64i8 = mul <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i8 = mul <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32i8 = mul <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V64i8 = mul <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'mul_constpow2'
@@ -118,10 +118,10 @@ define i32 @mul_constpow2() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i32 = mul <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, 16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = mul <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i16 = mul <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = mul <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i16 = mul <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i8 = mul <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i8 = mul <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32i8 = mul <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V64i8 = mul <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -451,12 +451,12 @@ define i32 @mul_constnegpow2() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i32 = mul <16 x i32> undef, <i32 -2, i32 -4, i32 -8, i32 -16, i32 -32, i32 -64, i32 -128, i32 -256, i32 -2, i32 -4, i32 -8, i32 -16, i32 -32, i32 -64, i32 -128, i32 -256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, -16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = mul <8 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i16 = mul <32 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32i16 = mul <32 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, -16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i8 = mul <32 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V64i8 = mul <64 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32i8 = mul <32 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V64i8 = mul <64 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'mul_constnegpow2'
@@ -470,10 +470,10 @@ define i32 @mul_constnegpow2() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i32 = mul <16 x i32> undef, <i32 -2, i32 -4, i32 -8, i32 -16, i32 -32, i32 -64, i32 -128, i32 -256, i32 -2, i32 -4, i32 -8, i32 -16, i32 -32, i32 -64, i32 -128, i32 -256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, -16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = mul <8 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i16 = mul <32 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, -16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32i8 = mul <32 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V64i8 = mul <64 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -489,10 +489,10 @@ define i32 @mul_constnegpow2() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i32 = mul <16 x i32> undef, <i32 -2, i32 -4, i32 -8, i32 -16, i32 -32, i32 -64, i32 -128, i32 -256, i32 -2, i32 -4, i32 -8, i32 -16, i32 -32, i32 -64, i32 -128, i32 -256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, -16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = mul <8 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32i16 = mul <32 x i16> undef, <i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256, i16 -2, i16 -4, i16 -8, i16 -16, i16 -32, i16 -64, i16 -128, i16 -256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, -16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32i8 = mul <32 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V64i8 = mul <64 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -627,12 +627,12 @@ define i32 @mul_uniformconstnegpow2() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i32 = mul <16 x i32> undef, <i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, -16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = mul <8 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i16 = mul <32 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32i16 = mul <32 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, -16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i8 = mul <32 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V64i8 = mul <64 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32i8 = mul <32 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V64i8 = mul <64 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'mul_uniformconstnegpow2'
@@ -646,10 +646,10 @@ define i32 @mul_uniformconstnegpow2() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i32 = mul <16 x i32> undef, <i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, -16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = mul <8 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i16 = mul <32 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, -16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32i8 = mul <32 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V64i8 = mul <64 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -665,10 +665,10 @@ define i32 @mul_uniformconstnegpow2() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i32 = mul <16 x i32> undef, <i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16, i32 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = mul i16 undef, -16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = mul <8 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = mul <16 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32i16 = mul <32 x i16> undef, <i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = mul i8 undef, -16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i8 = mul <16 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32i8 = mul <32 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V64i8 = mul <64 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef

diff --git a/llvm/test/Analysis/CostModel/X86/mul32.ll b/llvm/test/Analysis/CostModel/X86/mul32.ll
index 805c683..d50fc41 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/mul32.ll
+++ b/llvm/test/Analysis/CostModel/X86/mul32.ll
@@ -4,9 +4,9 @@
 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE42
 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+avx | FileCheck %s --check-prefixes=AVX1
 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512
-; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=AVX512
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
+; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=AVX512,AVX512F
 ;
 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mcpu=slm | FileCheck %s --check-prefixes=SLM
 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mcpu=goldmont | FileCheck %s --check-prefixes=GLM
@@ -268,29 +268,47 @@ define void @mul_zext_vXi8(<4 x i8> %a4, <4 x i8> %b4, <8 x i8> %a8, <8 x i8> %b
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %xa64 = zext <64 x i8> %a64 to <64 x i32>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %xb64 = zext <64 x i8> %b64 to <64 x i32>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
-; AVX512-LABEL: 'mul_zext_vXi8'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = zext <4 x i8> %a4 to <4 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = zext <4 x i8> %b4 to <4 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = zext <8 x i8> %a8 to <8 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = zext <8 x i8> %b8 to <8 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = zext <16 x i8> %a16 to <16 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = zext <16 x i8> %b16 to <16 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = zext <32 x i8> %a32 to <32 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = zext <32 x i8> %b32 to <32 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xa64 = zext <64 x i8> %a64 to <64 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xb64 = zext <64 x i8> %b64 to <64 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; AVX512F-LABEL: 'mul_zext_vXi8'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = zext <4 x i8> %a4 to <4 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = zext <4 x i8> %b4 to <4 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = zext <8 x i8> %a8 to <8 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = zext <8 x i8> %b8 to <8 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = zext <16 x i8> %a16 to <16 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = zext <16 x i8> %b16 to <16 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = zext <32 x i8> %a32 to <32 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = zext <32 x i8> %b32 to <32 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xa64 = zext <64 x i8> %a64 to <64 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xb64 = zext <64 x i8> %b64 to <64 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; AVX512BW-LABEL: 'mul_zext_vXi8'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = zext <4 x i8> %a4 to <4 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = zext <4 x i8> %b4 to <4 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = zext <8 x i8> %a8 to <8 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = zext <8 x i8> %b8 to <8 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = zext <16 x i8> %a16 to <16 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = zext <16 x i8> %b16 to <16 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = zext <32 x i8> %a32 to <32 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = zext <32 x i8> %b32 to <32 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xa64 = zext <64 x i8> %a64 to <64 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xb64 = zext <64 x i8> %b64 to <64 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; SLM-LABEL: 'mul_zext_vXi8'
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = zext <4 x i8> %a4 to <4 x i32>
@@ -431,29 +449,47 @@ define void @mul_sext_zext_vXi8(<4 x i8> %a4, <4 x i8> %b4, <8 x i8> %a8, <8 x i
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %xa64 = sext <64 x i8> %a64 to <64 x i32>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %xb64 = zext <64 x i8> %b64 to <64 x i32>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
-; AVX512-LABEL: 'mul_sext_zext_vXi8'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = sext <4 x i8> %a4 to <4 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = zext <4 x i8> %b4 to <4 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = sext <8 x i8> %a8 to <8 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = zext <8 x i8> %b8 to <8 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = sext <16 x i8> %a16 to <16 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = zext <16 x i8> %b16 to <16 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = sext <32 x i8> %a32 to <32 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = zext <32 x i8> %b32 to <32 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xa64 = sext <64 x i8> %a64 to <64 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xb64 = zext <64 x i8> %b64 to <64 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; AVX512F-LABEL: 'mul_sext_zext_vXi8'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = sext <4 x i8> %a4 to <4 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = zext <4 x i8> %b4 to <4 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = sext <8 x i8> %a8 to <8 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = zext <8 x i8> %b8 to <8 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = sext <16 x i8> %a16 to <16 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = zext <16 x i8> %b16 to <16 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = sext <32 x i8> %a32 to <32 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = zext <32 x i8> %b32 to <32 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xa64 = sext <64 x i8> %a64 to <64 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xb64 = zext <64 x i8> %b64 to <64 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; AVX512BW-LABEL: 'mul_sext_zext_vXi8'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = sext <4 x i8> %a4 to <4 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = zext <4 x i8> %b4 to <4 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = sext <8 x i8> %a8 to <8 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = zext <8 x i8> %b8 to <8 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = sext <16 x i8> %a16 to <16 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = zext <16 x i8> %b16 to <16 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = sext <32 x i8> %a32 to <32 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = zext <32 x i8> %b32 to <32 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xa64 = sext <64 x i8> %a64 to <64 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %xb64 = zext <64 x i8> %b64 to <64 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; SLM-LABEL: 'mul_sext_zext_vXi8'
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = sext <4 x i8> %a4 to <4 x i32>
@@ -598,29 +634,47 @@ define void @mul_sext_vXi16(<4 x i16> %a4, <4 x i16> %b4, <8 x i16> %a8, <8 x i1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %xa64 = sext <64 x i16> %a64 to <64 x i32>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %xb64 = sext <64 x i16> %b64 to <64 x i32>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
-; AVX512-LABEL: 'mul_sext_vXi16'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = sext <4 x i16> %a4 to <4 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = sext <4 x i16> %b4 to <4 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = sext <8 x i16> %a8 to <8 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = sext <8 x i16> %b8 to <8 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = sext <16 x i16> %a16 to <16 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = sext <16 x i16> %b16 to <16 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = sext <32 x i16> %a32 to <32 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = sext <32 x i16> %b32 to <32 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %xa64 = sext <64 x i16> %a64 to <64 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %xb64 = sext <64 x i16> %b64 to <64 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; AVX512F-LABEL: 'mul_sext_vXi16'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = sext <4 x i16> %a4 to <4 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = sext <4 x i16> %b4 to <4 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = sext <8 x i16> %a8 to <8 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = sext <8 x i16> %b8 to <8 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = sext <16 x i16> %a16 to <16 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = sext <16 x i16> %b16 to <16 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = sext <32 x i16> %a32 to <32 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = sext <32 x i16> %b32 to <32 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %xa64 = sext <64 x i16> %a64 to <64 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %xb64 = sext <64 x i16> %b64 to <64 x i32>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; AVX512BW-LABEL: 'mul_sext_vXi16'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = sext <4 x i16> %a4 to <4 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb4 = sext <4 x i16> %b4 to <4 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa8 = sext <8 x i16> %a8 to <8 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb8 = sext <8 x i16> %b8 to <8 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa16 = sext <16 x i16> %a16 to <16 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xb16 = sext <16 x i16> %b16 to <16 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xa32 = sext <32 x i16> %a32 to <32 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %xb32 = sext <32 x i16> %b32 to <32 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %xa64 = sext <64 x i16> %a64 to <64 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %xb64 = sext <64 x i16> %b64 to <64 x i32>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res4 = mul <4 x i32> %xa4, %xb4
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res8 = mul <8 x i32> %xa8, %xb8
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res16 = mul <16 x i32> %xa16, %xb16
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res32 = mul <32 x i32> %xa32, %xb32
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res64 = mul <64 x i32> %xa64, %xb64
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; SLM-LABEL: 'mul_sext_vXi16'
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %xa4 = sext <4 x i16> %a4 to <4 x i32>

diff --git a/llvm/test/Analysis/CostModel/X86/reduce-mul.ll b/llvm/test/Analysis/CostModel/X86/reduce-mul.ll
index 4e6d020..294345d 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/reduce-mul.ll
+++ b/llvm/test/Analysis/CostModel/X86/reduce-mul.ll
@@ -162,8 +162,8 @@ define i32 @reduce_i16(i32 %arg) {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4 = call i16 @llvm.vector.reduce.mul.v4i16(<4 x i16> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8 = call i16 @llvm.vector.reduce.mul.v8i16(<8 x i16> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16 = call i16 @llvm.vector.reduce.mul.v16i16(<16 x i16> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32 = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V64 = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32 = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'reduce_i16'
@@ -171,8 +171,8 @@ define i32 @reduce_i16(i32 %arg) {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4 = call i16 @llvm.vector.reduce.mul.v4i16(<4 x i16> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8 = call i16 @llvm.vector.reduce.mul.v8i16(<8 x i16> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16 = call i16 @llvm.vector.reduce.mul.v16i16(<16 x i16> undef)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32 = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> undef)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V64 = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> undef)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32 = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> undef)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64 = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512BW-LABEL: 'reduce_i16'
@@ -180,8 +180,8 @@ define i32 @reduce_i16(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4 = call i16 @llvm.vector.reduce.mul.v4i16(<4 x i16> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8 = call i16 @llvm.vector.reduce.mul.v8i16(<8 x i16> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16 = call i16 @llvm.vector.reduce.mul.v16i16(<16 x i16> undef)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32 = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> undef)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V64 = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> undef)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32 = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> undef)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V64 = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512DQ-LABEL: 'reduce_i16'
@@ -189,8 +189,8 @@ define i32 @reduce_i16(i32 %arg) {
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4 = call i16 @llvm.vector.reduce.mul.v4i16(<4 x i16> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8 = call i16 @llvm.vector.reduce.mul.v8i16(<8 x i16> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16 = call i16 @llvm.vector.reduce.mul.v16i16(<16 x i16> undef)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32 = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> undef)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V64 = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> undef)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32 = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> undef)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64 = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
   %V2  = call i16 @llvm.vector.reduce.mul.v2i16(<2 x i16> undef)
@@ -228,9 +228,9 @@ define i32 @reduce_i8(i32 %arg) {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4 = call i8 @llvm.vector.reduce.mul.v4i8(<4 x i8> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8 = call i8 @llvm.vector.reduce.mul.v8i8(<8 x i8> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16 = call i8 @llvm.vector.reduce.mul.v16i8(<16 x i8> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V32 = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V64 = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> undef)
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V128 = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V32 = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V64 = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> undef)
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 43 for instruction: %V128 = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> undef)
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'reduce_i8'
@@ -238,9 +238,9 @@ define i32 @reduce_i8(i32 %arg) {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4 = call i8 @llvm.vector.reduce.mul.v4i8(<4 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8 = call i8 @llvm.vector.reduce.mul.v8i8(<8 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16 = call i8 @llvm.vector.reduce.mul.v16i8(<16 x i8> undef)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32 = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> undef)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V64 = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> undef)
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> undef)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V32 = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> undef)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V64 = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> undef)
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V128 = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> undef)
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512BW-LABEL: 'reduce_i8'
@@ -248,9 +248,9 @@ define i32 @reduce_i8(i32 %arg) {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4 = call i8 @llvm.vector.reduce.mul.v4i8(<4 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8 = call i8 @llvm.vector.reduce.mul.v8i8(<8 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16 = call i8 @llvm.vector.reduce.mul.v16i8(<16 x i8> undef)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32 = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> undef)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> undef)
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V128 = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> undef)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32 = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> undef)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V64 = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> undef)
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V128 = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> undef)
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512DQ-LABEL: 'reduce_i8'
@@ -258,9 +258,9 @@ define i32 @reduce_i8(i32 %arg) {
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4 = call i8 @llvm.vector.reduce.mul.v4i8(<4 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8 = call i8 @llvm.vector.reduce.mul.v8i8(<8 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16 = call i8 @llvm.vector.reduce.mul.v16i8(<16 x i8> undef)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32 = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> undef)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V64 = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> undef)
-; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> undef)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V32 = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> undef)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V64 = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> undef)
+; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V128 = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> undef)
 ; AVX512DQ-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
   %V2   = call i8 @llvm.vector.reduce.mul.v2i8(<2 x i8> undef)

diff --git a/llvm/test/Analysis/CostModel/X86/rem-codesize.ll b/llvm/test/Analysis/CostModel/X86/rem-codesize.ll
index 70746a6..a80e68a 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/rem-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/rem-codesize.ll
@@ -1,14 +1,14 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512BW
 ;
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SSE,SLM
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=btver2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
 
 define i32 @srem() {
@@ -270,68 +270,87 @@ define i32 @urem_uniformconst() {
 }
 
 define i32 @srem_constpow2() {
-; SSE-LABEL: 'srem_constpow2'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'srem_constpow2'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 43 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'srem_constpow2'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 51 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX1-LABEL: 'srem_constpow2'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 45 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX2-LABEL: 'srem_constpow2'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'srem_constpow2'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
@@ -348,9 +367,9 @@ define i32 @srem_constpow2() {
 ;
 ; AVX512BW-LABEL: 'srem_constpow2'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
@@ -362,9 +381,28 @@ define i32 @srem_constpow2() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; SLM-LABEL: 'srem_constpow2'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = srem i64 undef, 16
   %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
   %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
@@ -470,68 +508,87 @@ define i32 @urem_constpow2() {
 }
 
 define i32 @srem_uniformconstpow2() {
-; SSE-LABEL: 'srem_uniformconstpow2'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'srem_uniformconstpow2'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 43 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'srem_uniformconstpow2'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 51 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX1-LABEL: 'srem_uniformconstpow2'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 45 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX2-LABEL: 'srem_uniformconstpow2'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'srem_uniformconstpow2'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
@@ -548,9 +605,9 @@ define i32 @srem_uniformconstpow2() {
 ;
 ; AVX512BW-LABEL: 'srem_uniformconstpow2'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
@@ -562,9 +619,28 @@ define i32 @srem_uniformconstpow2() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; SLM-LABEL: 'srem_uniformconstpow2'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = srem i64 undef, 16
   %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
   %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>

diff --git a/llvm/test/Analysis/CostModel/X86/rem-latency.ll b/llvm/test/Analysis/CostModel/X86/rem-latency.ll
index 606713f..f1bb531 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/rem-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/rem-latency.ll
@@ -1,14 +1,14 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512BW
 ;
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SLM
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SSE,SLM
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=btver2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
 
 define i32 @srem() {
@@ -270,118 +270,137 @@ define i32 @urem_uniformconst() {
 }
 
 define i32 @srem_constpow2() {
-; SSE-LABEL: 'srem_constpow2'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'srem_constpow2'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 37 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'srem_constpow2'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 49 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX1-LABEL: 'srem_constpow2'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 41 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 45 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX2-LABEL: 'srem_constpow2'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'srem_constpow2'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX512BW-LABEL: 'srem_constpow2'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; SLM-LABEL: 'srem_constpow2'
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 55 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 107 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 49 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %I64 = srem i64 undef, 16
@@ -465,25 +484,6 @@ define i32 @urem_constpow2() {
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64i8 = urem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
-; SLM-LABEL: 'urem_constpow2'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = urem i64 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2i64 = urem <2 x i64> undef, <i64 8, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i64 = urem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8i64 = urem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = urem i32 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4i32 = urem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i32 = urem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16i32 = urem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = urem i16 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = urem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = urem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32i16 = urem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = urem i8 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i8 = urem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i8 = urem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64i8 = urem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
-;
   %I64 = urem i64 undef, 16
   %V2i64 = urem <2 x i64> undef, <i64 8, i64 16>
   %V4i64 = urem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
@@ -508,118 +508,137 @@ define i32 @urem_constpow2() {
 }
 
 define i32 @srem_uniformconstpow2() {
-; SSE-LABEL: 'srem_uniformconstpow2'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'srem_uniformconstpow2'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 37 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'srem_uniformconstpow2'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 49 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX1-LABEL: 'srem_uniformconstpow2'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 41 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 45 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX2-LABEL: 'srem_uniformconstpow2'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'srem_uniformconstpow2'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX512BW-LABEL: 'srem_uniformconstpow2'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; SLM-LABEL: 'srem_uniformconstpow2'
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 55 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 107 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 49 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %I64 = srem i64 undef, 16
@@ -703,25 +722,6 @@ define i32 @urem_uniformconstpow2() {
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64i8 = urem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
-; SLM-LABEL: 'urem_uniformconstpow2'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = urem i64 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2i64 = urem <2 x i64> undef, <i64 16, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i64 = urem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8i64 = urem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = urem i32 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4i32 = urem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i32 = urem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16i32 = urem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = urem i16 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8i16 = urem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = urem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32i16 = urem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = urem i8 undef, 16
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16i8 = urem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i8 = urem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64i8 = urem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
-;
   %I64 = urem i64 undef, 16
   %V2i64 = urem <2 x i64> undef, <i64 16, i64 16>
   %V4i64 = urem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>

diff --git a/llvm/test/Analysis/CostModel/X86/rem-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/rem-sizelatency.ll
index b3cfc15..3b3753c 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/rem-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/rem-sizelatency.ll
@@ -1,14 +1,14 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+ssse3 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX512,AVX512BW
 ;
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=slm | FileCheck %s --check-prefixes=CHECK,SSE,SLM
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=goldmont | FileCheck %s --check-prefixes=CHECK,SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.8.0 -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=btver2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
 
 define i32 @srem() {
@@ -270,72 +270,91 @@ define i32 @urem_uniformconst() {
 }
 
 define i32 @srem_constpow2() {
-; SSE-LABEL: 'srem_constpow2'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'srem_constpow2'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 51 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'srem_constpow2'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX1-LABEL: 'srem_constpow2'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 35 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 67 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 39 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 53 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX2-LABEL: 'srem_constpow2'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 37 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'srem_constpow2'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
@@ -348,13 +367,13 @@ define i32 @srem_constpow2() {
 ;
 ; AVX512BW-LABEL: 'srem_constpow2'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
@@ -362,9 +381,28 @@ define i32 @srem_constpow2() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; SLM-LABEL: 'srem_constpow2'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 55 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = srem i64 undef, 16
   %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
   %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
@@ -470,72 +508,91 @@ define i32 @urem_constpow2() {
 }
 
 define i32 @srem_uniformconstpow2() {
-; SSE-LABEL: 'srem_uniformconstpow2'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE2-LABEL: 'srem_uniformconstpow2'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 51 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SSE42-LABEL: 'srem_uniformconstpow2'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX1-LABEL: 'srem_uniformconstpow2'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 35 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 67 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 39 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 53 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX2-LABEL: 'srem_uniformconstpow2'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 37 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'srem_uniformconstpow2'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
@@ -548,13 +605,13 @@ define i32 @srem_uniformconstpow2() {
 ;
 ; AVX512BW-LABEL: 'srem_uniformconstpow2'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
@@ -562,9 +619,28 @@ define i32 @srem_uniformconstpow2() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
+; SLM-LABEL: 'srem_uniformconstpow2'
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 55 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
+; SLM-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
   %I64 = srem i64 undef, 16
   %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
   %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>

diff --git a/llvm/test/Analysis/CostModel/X86/rem.ll b/llvm/test/Analysis/CostModel/X86/rem.ll
index 605bdf4..2154e97 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/rem.ll
+++ b/llvm/test/Analysis/CostModel/X86/rem.ll
@@ -889,12 +889,12 @@ define i32 @srem_constpow2() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 108 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'srem_constpow2'
@@ -908,10 +908,10 @@ define i32 @srem_constpow2() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 55 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 113 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -927,10 +927,10 @@ define i32 @srem_constpow2() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256, i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256, i16 2, i16 4, i16 8, i16 16, i16 32, i16 64, i16 128, i16 256>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V16i8 = srem <16 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %V32i8 = srem <32 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 60 for instruction: %V64i8 = srem <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1184,12 +1184,12 @@ define i32 @srem_uniformconstpow2() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 52 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512F-LABEL: 'srem_uniformconstpow2'
@@ -1203,10 +1203,10 @@ define i32 @srem_uniformconstpow2() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 53 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
@@ -1222,10 +1222,10 @@ define i32 @srem_uniformconstpow2() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = srem <16 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I16 = srem i16 undef, 16
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = srem <8 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i16 = srem <16 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32i16 = srem <32 x i16> undef, <i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %I8 = srem i8 undef, 16
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i8 = srem <16 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V32i8 = srem <32 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %V64i8 = srem <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef

diff --git a/llvm/test/Analysis/CostModel/X86/vshift-shl-codesize.ll b/llvm/test/Analysis/CostModel/X86/vshift-shl-codesize.ll
index eaec9d2..8c1904b 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/vshift-shl-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/vshift-shl-codesize.ll
@@ -754,27 +754,99 @@ define <8 x i64> @constant_shift_v8i64(<8 x i64> %a) {
 }
 
 define <4 x i32> @constant_shift_v4i32(<4 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v4i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v4i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; AVX-LABEL: 'constant_shift_v4i32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; XOP-LABEL: 'constant_shift_v4i32'
+; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v4i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
 ;
   %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
   ret <4 x i32> %shift
 }
 
 define <8 x i32> @constant_shift_v8i32(<8 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v8i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v8i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v8i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX1-LABEL: 'constant_shift_v8i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX2-LABEL: 'constant_shift_v8i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v8i32'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v8i32'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v8i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v8i32'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
 ;
   %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
   ret <8 x i32> %shift
 }
 
 define <16 x i32> @constant_shift_v16i32(<16 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v16i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v16i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v16i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX1-LABEL: 'constant_shift_v16i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX2-LABEL: 'constant_shift_v16i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v16i32'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v16i32'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v16i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v16i32'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
 ;
   %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
   ret <16 x i32> %shift
@@ -790,18 +862,74 @@ define <8 x i16> @constant_shift_v8i16(<8 x i16> %a) {
 }
 
 define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) {
-; CHECK-LABEL: 'constant_shift_v16i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+; SSE2-LABEL: 'constant_shift_v16i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; SSE42-LABEL: 'constant_shift_v16i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX1-LABEL: 'constant_shift_v16i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX2-LABEL: 'constant_shift_v16i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v16i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v16i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX512-LABEL: 'constant_shift_v16i16'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v16i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
 ;
   %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
   ret <16 x i16> %shift
 }
 
 define <32 x i16> @constant_shift_v32i16(<32 x i16> %a) {
-; CHECK-LABEL: 'constant_shift_v32i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+; SSE2-LABEL: 'constant_shift_v32i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; SSE42-LABEL: 'constant_shift_v32i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX1-LABEL: 'constant_shift_v32i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX2-LABEL: 'constant_shift_v32i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v32i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v32i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX512-LABEL: 'constant_shift_v32i16'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v32i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
 ;
   %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
   ret <32 x i16> %shift
@@ -911,9 +1039,37 @@ define <16 x i16> @splatconstant_shift_v16i16(<16 x i16> %a) {
 }
 
 define <32 x i16> @splatconstant_shift_v32i16(<32 x i16> %a) {
-; CHECK-LABEL: 'splatconstant_shift_v32i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+; SSE2-LABEL: 'splatconstant_shift_v32i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; SSE42-LABEL: 'splatconstant_shift_v32i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX1-LABEL: 'splatconstant_shift_v32i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX2-LABEL: 'splatconstant_shift_v32i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX1-LABEL: 'splatconstant_shift_v32i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX2-LABEL: 'splatconstant_shift_v32i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX512-LABEL: 'splatconstant_shift_v32i16'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; BTVER2-LABEL: 'splatconstant_shift_v32i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
 ;
   %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ret <32 x i16> %shift

diff --git a/llvm/test/Analysis/CostModel/X86/vshift-shl-cost-inseltpoison.ll b/llvm/test/Analysis/CostModel/X86/vshift-shl-cost-inseltpoison.ll
index 6bc84b0..d999e63 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/vshift-shl-cost-inseltpoison.ll
+++ b/llvm/test/Analysis/CostModel/X86/vshift-shl-cost-inseltpoison.ll
@@ -1387,7 +1387,7 @@ define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
 ; AVX2-LABEL: 'constant_shift_v16i16'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
 ; XOPAVX1-LABEL: 'constant_shift_v16i16'
@@ -1395,12 +1395,24 @@ define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) {
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
 ; XOPAVX2-LABEL: 'constant_shift_v16i16'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
-; AVX512-LABEL: 'constant_shift_v16i16'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
+; AVX512F-LABEL: 'constant_shift_v16i16'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
+;
+; AVX512BW-LABEL: 'constant_shift_v16i16'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
+;
+; AVX512VL-LABEL: 'constant_shift_v16i16'
+; AVX512VL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512VL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
+;
+; AVX512BWVL-LABEL: 'constant_shift_v16i16'
+; AVX512BWVL-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512BWVL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
 ; BTVER2-LABEL: 'constant_shift_v16i16'
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
@@ -1420,7 +1432,7 @@ define <32 x i16> @constant_shift_v32i16(<32 x i16> %a) {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shift
 ;
 ; AVX2-LABEL: 'constant_shift_v32i16'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shift
 ;
 ; XOPAVX1-LABEL: 'constant_shift_v32i16'
@@ -1428,7 +1440,7 @@ define <32 x i16> @constant_shift_v32i16(<32 x i16> %a) {
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shift
 ;
 ; XOPAVX2-LABEL: 'constant_shift_v32i16'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shift
 ;
 ; AVX512F-LABEL: 'constant_shift_v32i16'
@@ -2066,7 +2078,7 @@ define <16 x i16> @test6(<16 x i16> %a) {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
 ; AVX2-LABEL: 'test6'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
 ; XOPAVX1-LABEL: 'test6'
@@ -2074,12 +2086,24 @@ define <16 x i16> @test6(<16 x i16> %a) {
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
 ; XOPAVX2-LABEL: 'test6'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
-; AVX512-LABEL: 'test6'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
+; AVX512F-LABEL: 'test6'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
+;
+; AVX512BW-LABEL: 'test6'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
+;
+; AVX512VL-LABEL: 'test6'
+; AVX512VL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX512VL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
+;
+; AVX512BWVL-LABEL: 'test6'
+; AVX512BWVL-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX512BWVL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
 ; BTVER2-LABEL: 'test6'
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
@@ -2180,7 +2204,7 @@ define <32 x i16> @test9(<32 x i16> %a) {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shl
 ;
 ; AVX2-LABEL: 'test9'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <32 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shl = shl <32 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shl
 ;
 ; XOPAVX1-LABEL: 'test9'
@@ -2188,7 +2212,7 @@ define <32 x i16> @test9(<32 x i16> %a) {
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shl
 ;
 ; XOPAVX2-LABEL: 'test9'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <32 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shl = shl <32 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shl
 ;
 ; AVX512F-LABEL: 'test9'

diff --git a/llvm/test/Analysis/CostModel/X86/vshift-shl-cost.ll b/llvm/test/Analysis/CostModel/X86/vshift-shl-cost.ll
index be56c19..7fb5b36 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/vshift-shl-cost.ll
+++ b/llvm/test/Analysis/CostModel/X86/vshift-shl-cost.ll
@@ -1387,7 +1387,7 @@ define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
 ; AVX2-LABEL: 'constant_shift_v16i16'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
 ; XOPAVX1-LABEL: 'constant_shift_v16i16'
@@ -1395,12 +1395,24 @@ define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) {
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
 ; XOPAVX2-LABEL: 'constant_shift_v16i16'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
-; AVX512-LABEL: 'constant_shift_v16i16'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
+; AVX512F-LABEL: 'constant_shift_v16i16'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
+;
+; AVX512BW-LABEL: 'constant_shift_v16i16'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
+;
+; AVX512VL-LABEL: 'constant_shift_v16i16'
+; AVX512VL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512VL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
+;
+; AVX512BWVL-LABEL: 'constant_shift_v16i16'
+; AVX512BWVL-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512BWVL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shift
 ;
 ; BTVER2-LABEL: 'constant_shift_v16i16'
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
@@ -1420,7 +1432,7 @@ define <32 x i16> @constant_shift_v32i16(<32 x i16> %a) {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shift
 ;
 ; AVX2-LABEL: 'constant_shift_v32i16'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shift
 ;
 ; XOPAVX1-LABEL: 'constant_shift_v32i16'
@@ -1428,7 +1440,7 @@ define <32 x i16> @constant_shift_v32i16(<32 x i16> %a) {
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shift
 ;
 ; XOPAVX2-LABEL: 'constant_shift_v32i16'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shift
 ;
 ; AVX512F-LABEL: 'constant_shift_v32i16'
@@ -2066,7 +2078,7 @@ define <16 x i16> @test6(<16 x i16> %a) {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
 ; AVX2-LABEL: 'test6'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
 ; XOPAVX1-LABEL: 'test6'
@@ -2074,12 +2086,24 @@ define <16 x i16> @test6(<16 x i16> %a) {
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
 ; XOPAVX2-LABEL: 'test6'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
-; AVX512-LABEL: 'test6'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
+; AVX512F-LABEL: 'test6'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
+;
+; AVX512BW-LABEL: 'test6'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
+;
+; AVX512VL-LABEL: 'test6'
+; AVX512VL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX512VL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
+;
+; AVX512BWVL-LABEL: 'test6'
+; AVX512BWVL-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX512BWVL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %shl
 ;
 ; BTVER2-LABEL: 'test6'
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shl = shl <16 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
@@ -2180,7 +2204,7 @@ define <32 x i16> @test9(<32 x i16> %a) {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shl
 ;
 ; AVX2-LABEL: 'test9'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <32 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shl = shl <32 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shl
 ;
 ; XOPAVX1-LABEL: 'test9'
@@ -2188,7 +2212,7 @@ define <32 x i16> @test9(<32 x i16> %a) {
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shl
 ;
 ; XOPAVX2-LABEL: 'test9'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shl = shl <32 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shl = shl <32 x i16> %a, <i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11, i16 1, i16 1, i16 2, i16 3, i16 7, i16 0, i16 9, i16 11>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i16> %shl
 ;
 ; AVX512F-LABEL: 'test9'

diff --git a/llvm/test/Analysis/CostModel/X86/vshift-shl-latency.ll b/llvm/test/Analysis/CostModel/X86/vshift-shl-latency.ll
index c273233..3d3cd00 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/vshift-shl-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/vshift-shl-latency.ll
@@ -754,27 +754,111 @@ define <8 x i64> @constant_shift_v8i64(<8 x i64> %a) {
 }
 
 define <4 x i32> @constant_shift_v4i32(<4 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v4i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v4i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; AVX1-LABEL: 'constant_shift_v4i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; AVX2-LABEL: 'constant_shift_v4i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v4i32'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v4i32'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v4i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v4i32'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
 ;
   %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
   ret <4 x i32> %shift
 }
 
 define <8 x i32> @constant_shift_v8i32(<8 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v8i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v8i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v8i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX1-LABEL: 'constant_shift_v8i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX2-LABEL: 'constant_shift_v8i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v8i32'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v8i32'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v8i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v8i32'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
 ;
   %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
   ret <8 x i32> %shift
 }
 
 define <16 x i32> @constant_shift_v16i32(<16 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v16i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v16i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v16i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 44 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX1-LABEL: 'constant_shift_v16i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX2-LABEL: 'constant_shift_v16i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v16i32'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v16i32'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v16i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v16i32'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
 ;
   %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
   ret <16 x i32> %shift
@@ -782,7 +866,7 @@ define <16 x i32> @constant_shift_v16i32(<16 x i32> %a) {
 
 define <8 x i16> @constant_shift_v8i16(<8 x i16> %a) {
 ; CHECK-LABEL: 'constant_shift_v8i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <8 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <8 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i16> %shift
 ;
   %shift = shl <8 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
@@ -790,18 +874,70 @@ define <8 x i16> @constant_shift_v8i16(<8 x i16> %a) {
 }
 
 define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) {
-; CHECK-LABEL: 'constant_shift_v16i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+; SSE2-LABEL: 'constant_shift_v16i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; SSE42-LABEL: 'constant_shift_v16i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX1-LABEL: 'constant_shift_v16i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX2-LABEL: 'constant_shift_v16i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v16i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v16i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX512-LABEL: 'constant_shift_v16i16'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v16i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
 ;
   %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
   ret <16 x i16> %shift
 }
 
 define <32 x i16> @constant_shift_v32i16(<32 x i16> %a) {
-; CHECK-LABEL: 'constant_shift_v32i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+; SSE2-LABEL: 'constant_shift_v32i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; SSE42-LABEL: 'constant_shift_v32i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX1-LABEL: 'constant_shift_v32i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX2-LABEL: 'constant_shift_v32i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v32i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v32i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v32i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
 ;
   %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
   ret <32 x i16> %shift
@@ -902,18 +1038,70 @@ define <8 x i16> @splatconstant_shift_v8i16(<8 x i16> %a) {
 }
 
 define <16 x i16> @splatconstant_shift_v16i16(<16 x i16> %a) {
-; CHECK-LABEL: 'splatconstant_shift_v16i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+; SSE2-LABEL: 'splatconstant_shift_v16i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; SSE42-LABEL: 'splatconstant_shift_v16i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX1-LABEL: 'splatconstant_shift_v16i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX2-LABEL: 'splatconstant_shift_v16i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; XOPAVX1-LABEL: 'splatconstant_shift_v16i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; XOPAVX2-LABEL: 'splatconstant_shift_v16i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX512-LABEL: 'splatconstant_shift_v16i16'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; BTVER2-LABEL: 'splatconstant_shift_v16i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
 ;
   %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ret <16 x i16> %shift
 }
 
 define <32 x i16> @splatconstant_shift_v32i16(<32 x i16> %a) {
-; CHECK-LABEL: 'splatconstant_shift_v32i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+; SSE2-LABEL: 'splatconstant_shift_v32i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; SSE42-LABEL: 'splatconstant_shift_v32i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX1-LABEL: 'splatconstant_shift_v32i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX2-LABEL: 'splatconstant_shift_v32i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX1-LABEL: 'splatconstant_shift_v32i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX2-LABEL: 'splatconstant_shift_v32i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; BTVER2-LABEL: 'splatconstant_shift_v32i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
 ;
   %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ret <32 x i16> %shift

diff --git a/llvm/test/Analysis/CostModel/X86/vshift-shl-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/vshift-shl-sizelatency.ll
index ae9c30f..9247d1c 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/vshift-shl-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/vshift-shl-sizelatency.ll
@@ -754,27 +754,111 @@ define <8 x i64> @constant_shift_v8i64(<8 x i64> %a) {
 }
 
 define <4 x i32> @constant_shift_v4i32(<4 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v4i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v4i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; AVX1-LABEL: 'constant_shift_v4i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; AVX2-LABEL: 'constant_shift_v4i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v4i32'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v4i32'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v4i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v4i32'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <4 x i32> %shift
 ;
   %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
   ret <4 x i32> %shift
 }
 
 define <8 x i32> @constant_shift_v8i32(<8 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v8i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v8i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v8i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX1-LABEL: 'constant_shift_v8i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX2-LABEL: 'constant_shift_v8i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v8i32'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v8i32'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v8i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v8i32'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i32> %shift
 ;
   %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
   ret <8 x i32> %shift
 }
 
 define <16 x i32> @constant_shift_v16i32(<16 x i32> %a) {
-; CHECK-LABEL: 'constant_shift_v16i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+; SSE2-LABEL: 'constant_shift_v16i32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; SSE42-LABEL: 'constant_shift_v16i32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX1-LABEL: 'constant_shift_v16i32'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX2-LABEL: 'constant_shift_v16i32'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v16i32'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v16i32'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; AVX512-LABEL: 'constant_shift_v16i32'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v16i32'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i32> %shift
 ;
   %shift = shl <16 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
   ret <16 x i32> %shift
@@ -790,18 +874,74 @@ define <8 x i16> @constant_shift_v8i16(<8 x i16> %a) {
 }
 
 define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) {
-; CHECK-LABEL: 'constant_shift_v16i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+; SSE2-LABEL: 'constant_shift_v16i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; SSE42-LABEL: 'constant_shift_v16i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX1-LABEL: 'constant_shift_v16i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX2-LABEL: 'constant_shift_v16i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v16i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v16i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; AVX512-LABEL: 'constant_shift_v16i16'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v16i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <16 x i16> %shift
 ;
   %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
   ret <16 x i16> %shift
 }
 
 define <32 x i16> @constant_shift_v32i16(<32 x i16> %a) {
-; CHECK-LABEL: 'constant_shift_v32i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+; SSE2-LABEL: 'constant_shift_v32i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; SSE42-LABEL: 'constant_shift_v32i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX1-LABEL: 'constant_shift_v32i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX2-LABEL: 'constant_shift_v32i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX1-LABEL: 'constant_shift_v32i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX2-LABEL: 'constant_shift_v32i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX512-LABEL: 'constant_shift_v32i16'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; BTVER2-LABEL: 'constant_shift_v32i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
 ;
   %shift = shl <32 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
   ret <32 x i16> %shift
@@ -911,9 +1051,37 @@ define <16 x i16> @splatconstant_shift_v16i16(<16 x i16> %a) {
 }
 
 define <32 x i16> @splatconstant_shift_v32i16(<32 x i16> %a) {
-; CHECK-LABEL: 'splatconstant_shift_v32i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+; SSE2-LABEL: 'splatconstant_shift_v32i16'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; SSE42-LABEL: 'splatconstant_shift_v32i16'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX1-LABEL: 'splatconstant_shift_v32i16'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX2-LABEL: 'splatconstant_shift_v32i16'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX1-LABEL: 'splatconstant_shift_v32i16'
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; XOPAVX2-LABEL: 'splatconstant_shift_v32i16'
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; AVX512-LABEL: 'splatconstant_shift_v32i16'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
+;
+; BTVER2-LABEL: 'splatconstant_shift_v32i16'
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <32 x i16> %shift
 ;
   %shift = shl <32 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ret <32 x i16> %shift

diff --git a/llvm/test/tools/llvm-mca/X86/Znver1/resources-avx2.s b/llvm/test/tools/llvm-mca/X86/Znver1/resources-avx2.s
index 669e6f7..c77fe3c 100644 (file)
--- a/llvm/test/tools/llvm-mca/X86/Znver1/resources-avx2.s
+++ b/llvm/test/tools/llvm-mca/X86/Znver1/resources-avx2.s
@@ -588,10 +588,10 @@ vpxor           (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  1      10    0.50    *                   vphsubsw       (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  1      3     0.25                        vphsubw        %ymm0, %ymm1, %ymm2
 # CHECK-NEXT:  1      10    0.50    *                   vphsubw        (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  1      4     1.00                        vpmaddubsw     %ymm0, %ymm1, %ymm2
-# CHECK-NEXT:  1      11    1.00    *                   vpmaddubsw     (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  1      4     1.00                        vpmaddwd       %ymm0, %ymm1, %ymm2
-# CHECK-NEXT:  1      11    1.00    *                   vpmaddwd       (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  2      4     2.00                        vpmaddubsw     %ymm0, %ymm1, %ymm2
+# CHECK-NEXT:  2      11    2.00    *                   vpmaddubsw     (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  2      4     2.00                        vpmaddwd       %ymm0, %ymm1, %ymm2
+# CHECK-NEXT:  2      11    2.00    *                   vpmaddwd       (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  1      100   0.25    *                   vpmaskmovd     (%rax), %xmm0, %xmm2
 # CHECK-NEXT:  1      100   0.25    *                   vpmaskmovd     (%rax), %ymm0, %ymm2
 # CHECK-NEXT:  1      100   0.25    *      *            vpmaskmovd     %xmm0, %xmm1, (%rax)
@@ -649,20 +649,20 @@ vpxor           (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  3      8     2.00    *                   vpmovzxwd      (%rax), %ymm2
 # CHECK-NEXT:  3      1     2.00                        vpmovzxwq      %xmm0, %ymm2
 # CHECK-NEXT:  3      8     2.00    *                   vpmovzxwq      (%rax), %ymm2
-# CHECK-NEXT:  1      4     1.00                        vpmuldq        %ymm0, %ymm1, %ymm2
-# CHECK-NEXT:  1      11    1.00    *                   vpmuldq        (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  1      4     1.00                        vpmulhrsw      %ymm0, %ymm1, %ymm2
-# CHECK-NEXT:  1      11    1.00    *                   vpmulhrsw      (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  1      4     1.00                        vpmulhuw       %ymm0, %ymm1, %ymm2
-# CHECK-NEXT:  1      11    1.00    *                   vpmulhuw       (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  1      4     1.00                        vpmulhw        %ymm0, %ymm1, %ymm2
-# CHECK-NEXT:  1      11    1.00    *                   vpmulhw        (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  2      4     2.00                        vpmuldq        %ymm0, %ymm1, %ymm2
+# CHECK-NEXT:  2      11    2.00    *                   vpmuldq        (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  2      4     2.00                        vpmulhrsw      %ymm0, %ymm1, %ymm2
+# CHECK-NEXT:  2      11    2.00    *                   vpmulhrsw      (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  2      4     2.00                        vpmulhuw       %ymm0, %ymm1, %ymm2
+# CHECK-NEXT:  2      11    2.00    *                   vpmulhuw       (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  2      4     2.00                        vpmulhw        %ymm0, %ymm1, %ymm2
+# CHECK-NEXT:  2      11    2.00    *                   vpmulhw        (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  2      4     4.00                        vpmulld        %ymm0, %ymm1, %ymm2
 # CHECK-NEXT:  2      11    4.00    *                   vpmulld        (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  1      4     1.00                        vpmullw        %ymm0, %ymm1, %ymm2
-# CHECK-NEXT:  1      11    1.00    *                   vpmullw        (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  1      4     1.00                        vpmuludq       %ymm0, %ymm1, %ymm2
-# CHECK-NEXT:  1      11    1.00    *                   vpmuludq       (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  2      4     2.00                        vpmullw        %ymm0, %ymm1, %ymm2
+# CHECK-NEXT:  2      11    2.00    *                   vpmullw        (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  2      4     2.00                        vpmuludq       %ymm0, %ymm1, %ymm2
+# CHECK-NEXT:  2      11    2.00    *                   vpmuludq       (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  2      1     0.50                        vpor   %ymm0, %ymm1, %ymm2
 # CHECK-NEXT:  2      8     0.50    *                   vpor   (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  1      3     1.00                        vpsadbw        %ymm0, %ymm1, %ymm2
@@ -778,7 +778,7 @@ vpxor           (%rax), %ymm1, %ymm2
 
 # CHECK:      Resource pressure per iteration:
 # CHECK-NEXT: [0]    [1]    [2]    [3]    [4]    [5]    [6]    [7]    [8]    [9]    [10]   [11]
-# CHECK-NEXT: 67.00  67.00   -      -      -      -      -     100.67 183.17 164.50 64.67   -
+# CHECK-NEXT: 67.00  67.00   -      -      -      -      -     116.67 183.17 164.50 64.67   -
 
 # CHECK:      Resource pressure by instruction:
 # CHECK-NEXT: [0]    [1]    [2]    [3]    [4]    [5]    [6]    [7]    [8]    [9]    [10]   [11]   Instructions:
@@ -910,10 +910,10 @@ vpxor           (%rax), %ymm1, %ymm2
 # CHECK-NEXT: 0.50   0.50    -      -      -      -      -      -      -      -      -      -     vphsubsw     (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  -      -      -      -      -      -      -      -      -      -      -      -     vphsubw      %ymm0, %ymm1, %ymm2
 # CHECK-NEXT: 0.50   0.50    -      -      -      -      -      -      -      -      -      -     vphsubw      (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpmaddubsw   %ymm0, %ymm1, %ymm2
-# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     1.00    -      -      -      -     vpmaddubsw   (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpmaddwd     %ymm0, %ymm1, %ymm2
-# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     1.00    -      -      -      -     vpmaddwd     (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  -      -      -      -      -      -      -     2.00    -      -      -      -     vpmaddubsw   %ymm0, %ymm1, %ymm2
+# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     2.00    -      -      -      -     vpmaddubsw   (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  -      -      -      -      -      -      -     2.00    -      -      -      -     vpmaddwd     %ymm0, %ymm1, %ymm2
+# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     2.00    -      -      -      -     vpmaddwd     (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  -      -      -      -      -      -      -      -      -      -      -      -     vpmaskmovd   (%rax), %xmm0, %xmm2
 # CHECK-NEXT:  -      -      -      -      -      -      -      -      -      -      -      -     vpmaskmovd   (%rax), %ymm0, %ymm2
 # CHECK-NEXT:  -      -      -      -      -      -      -      -      -      -      -      -     vpmaskmovd   %xmm0, %xmm1, (%rax)
@@ -971,20 +971,20 @@ vpxor           (%rax), %ymm1, %ymm2
 # CHECK-NEXT: 0.50   0.50    -      -      -      -      -      -     2.00   2.00    -      -     vpmovzxwd    (%rax), %ymm2
 # CHECK-NEXT:  -      -      -      -      -      -      -      -     2.00   2.00    -      -     vpmovzxwq    %xmm0, %ymm2
 # CHECK-NEXT: 0.50   0.50    -      -      -      -      -      -     2.00   2.00    -      -     vpmovzxwq    (%rax), %ymm2
-# CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpmuldq      %ymm0, %ymm1, %ymm2
-# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     1.00    -      -      -      -     vpmuldq      (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpmulhrsw    %ymm0, %ymm1, %ymm2
-# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     1.00    -      -      -      -     vpmulhrsw    (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpmulhuw     %ymm0, %ymm1, %ymm2
-# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     1.00    -      -      -      -     vpmulhuw     (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpmulhw      %ymm0, %ymm1, %ymm2
-# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     1.00    -      -      -      -     vpmulhw      (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  -      -      -      -      -      -      -     2.00    -      -      -      -     vpmuldq      %ymm0, %ymm1, %ymm2
+# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     2.00    -      -      -      -     vpmuldq      (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  -      -      -      -      -      -      -     2.00    -      -      -      -     vpmulhrsw    %ymm0, %ymm1, %ymm2
+# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     2.00    -      -      -      -     vpmulhrsw    (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  -      -      -      -      -      -      -     2.00    -      -      -      -     vpmulhuw     %ymm0, %ymm1, %ymm2
+# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     2.00    -      -      -      -     vpmulhuw     (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  -      -      -      -      -      -      -     2.00    -      -      -      -     vpmulhw      %ymm0, %ymm1, %ymm2
+# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     2.00    -      -      -      -     vpmulhw      (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  -      -      -      -      -      -      -     4.00    -      -      -      -     vpmulld      %ymm0, %ymm1, %ymm2
 # CHECK-NEXT: 0.50   0.50    -      -      -      -      -     4.00    -      -      -      -     vpmulld      (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpmullw      %ymm0, %ymm1, %ymm2
-# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     1.00    -      -      -      -     vpmullw      (%rax), %ymm1, %ymm2
-# CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpmuludq     %ymm0, %ymm1, %ymm2
-# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     1.00    -      -      -      -     vpmuludq     (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  -      -      -      -      -      -      -     2.00    -      -      -      -     vpmullw      %ymm0, %ymm1, %ymm2
+# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     2.00    -      -      -      -     vpmullw      (%rax), %ymm1, %ymm2
+# CHECK-NEXT:  -      -      -      -      -      -      -     2.00    -      -      -      -     vpmuludq     %ymm0, %ymm1, %ymm2
+# CHECK-NEXT: 0.50   0.50    -      -      -      -      -     2.00    -      -      -      -     vpmuludq     (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  -      -      -      -      -      -      -     0.50   0.50   0.50   0.50    -     vpor %ymm0, %ymm1, %ymm2
 # CHECK-NEXT: 0.50   0.50    -      -      -      -      -     0.50   0.50   0.50   0.50    -     vpor (%rax), %ymm1, %ymm2
 # CHECK-NEXT:  -      -      -      -      -      -      -     1.00    -      -      -      -     vpsadbw      %ymm0, %ymm1, %ymm2