[CostModel][X86] Add vXi32 division by uniform constant costs (PR47476)

Other types can be handled in future patches but their uniform / non-uniform costs are more similar and don't appear to cause many vectorization issues.

diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index c917974..03f8be0 100644 (file)
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -321,6 +321,11 @@ int X86TTIImpl::getArithmeticInstrCost(unsigned Opcode, Type *Ty,
     { ISD::SHL,  MVT::v64i8,   4 }, // psllw + pand.
     { ISD::SRL,  MVT::v64i8,   4 }, // psrlw + pand.
     { ISD::SRA,  MVT::v64i8,   8 }, // psrlw, pand, pxor, psubb.
+
+    { ISD::SDIV, MVT::v16i32,  6 }, // pmuludq sequence
+    { ISD::SREM, MVT::v16i32,  8 }, // pmuludq+mul+sub sequence
+    { ISD::UDIV, MVT::v16i32,  5 }, // pmuludq sequence
+    { ISD::UREM, MVT::v16i32,  7 }, // pmuludq+mul+sub sequence
   };
 
   if (Op2Info == TargetTransformInfo::OK_UniformConstantValue &&
@@ -336,6 +341,11 @@ int X86TTIImpl::getArithmeticInstrCost(unsigned Opcode, Type *Ty,
     { ISD::SRA,  MVT::v32i8,   4 }, // psrlw, pand, pxor, psubb.
 
     { ISD::SRA,  MVT::v4i64,   4 }, // 2 x psrad + shuffle.
+
+    { ISD::SDIV, MVT::v8i32,   6 }, // pmuludq sequence
+    { ISD::SREM, MVT::v8i32,   8 }, // pmuludq+mul+sub sequence
+    { ISD::UDIV, MVT::v8i32,   5 }, // pmuludq sequence
+    { ISD::UREM, MVT::v8i32,   7 }, // pmuludq+mul+sub sequence
   };
 
   if (Op2Info == TargetTransformInfo::OK_UniformConstantValue &&
@@ -353,6 +363,15 @@ int X86TTIImpl::getArithmeticInstrCost(unsigned Opcode, Type *Ty,
     { ISD::SHL,  MVT::v32i8,   4+2 }, // 2*(psllw + pand) + split.
     { ISD::SRL,  MVT::v32i8,   4+2 }, // 2*(psrlw + pand) + split.
     { ISD::SRA,  MVT::v32i8,   8+2 }, // 2*(psrlw, pand, pxor, psubb) + split.
+
+    { ISD::SDIV, MVT::v8i32,  12+2 }, // 2*pmuludq sequence + split.
+    { ISD::SREM, MVT::v8i32,  16+2 }, // 2*pmuludq+mul+sub sequence + split.
+    { ISD::SDIV, MVT::v4i32,     6 }, // pmuludq sequence
+    { ISD::SREM, MVT::v4i32,     8 }, // pmuludq+mul+sub sequence
+    { ISD::UDIV, MVT::v8i32,  10+2 }, // 2*pmuludq sequence + split.
+    { ISD::UREM, MVT::v8i32,  14+2 }, // 2*pmuludq+mul+sub sequence + split.
+    { ISD::UDIV, MVT::v4i32,     5 }, // pmuludq sequence
+    { ISD::UREM, MVT::v4i32,     7 }, // pmuludq+mul+sub sequence
   };
 
   // XOP has faster vXi8 shifts.

diff --git a/llvm/test/Analysis/CostModel/X86/div.ll b/llvm/test/Analysis/CostModel/X86/div.ll
index fb3b705..4bead92 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/div.ll
+++ b/llvm/test/Analysis/CostModel/X86/div.ll
@@ -450,62 +450,24 @@ define i32 @udiv_const() {
 }
 
 define i32 @sdiv_uniformconst() {
-; SSE2-LABEL: 'sdiv_uniformconst'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sdiv i64 undef, 7
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 7, i64 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 76 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32i16 = sdiv <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = sdiv i8 undef, 7
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i8 = sdiv <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32i8 = sdiv <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'sdiv_uniformconst'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sdiv i64 undef, 7
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 7, i64 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 76 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32i16 = sdiv <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = sdiv i8 undef, 7
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i8 = sdiv <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32i8 = sdiv <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSE42-LABEL: 'sdiv_uniformconst'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sdiv i64 undef, 7
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 7, i64 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 60 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32i16 = sdiv <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = sdiv i8 undef, 7
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i8 = sdiv <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32i8 = sdiv <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; SSE-LABEL: 'sdiv_uniformconst'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sdiv i64 undef, 7
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 7, i64 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
+; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
+; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V32i16 = sdiv <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = sdiv i8 undef, 7
+; SSE-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i8 = sdiv <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V32i8 = sdiv <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX1-LABEL: 'sdiv_uniformconst'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sdiv i64 undef, 7
@@ -513,9 +475,9 @@ define i32 @sdiv_uniformconst() {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -532,9 +494,9 @@ define i32 @sdiv_uniformconst() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -551,9 +513,9 @@ define i32 @sdiv_uniformconst() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -570,9 +532,9 @@ define i32 @sdiv_uniformconst() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -589,9 +551,9 @@ define i32 @sdiv_uniformconst() {
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; SLM-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 60 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -608,9 +570,9 @@ define i32 @sdiv_uniformconst() {
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = sdiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = sdiv i32 undef, 7
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4i32 = sdiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i32 = sdiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V16i32 = sdiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = sdiv i16 undef, 7
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = sdiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i16 = sdiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -651,9 +613,9 @@ define i32 @udiv_uniformconst() {
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = udiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = udiv i32 undef, 7
-; SSE-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 60 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = udiv i16 undef, 7
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = udiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i16 = udiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -670,9 +632,9 @@ define i32 @udiv_uniformconst() {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = udiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = udiv i32 undef, 7
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = udiv i16 undef, 7
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = udiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i16 = udiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -689,9 +651,9 @@ define i32 @udiv_uniformconst() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = udiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = udiv i32 undef, 7
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = udiv i16 undef, 7
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = udiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = udiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -708,9 +670,9 @@ define i32 @udiv_uniformconst() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = udiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = udiv i32 undef, 7
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = udiv i16 undef, 7
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = udiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = udiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -727,9 +689,9 @@ define i32 @udiv_uniformconst() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = udiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = udiv i32 undef, 7
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = udiv i16 undef, 7
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = udiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16i16 = udiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -746,9 +708,9 @@ define i32 @udiv_uniformconst() {
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = udiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = udiv i32 undef, 7
-; SLM-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 60 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = udiv i16 undef, 7
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = udiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16i16 = udiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -765,9 +727,9 @@ define i32 @udiv_uniformconst() {
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = udiv <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = udiv i32 undef, 7
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4i32 = udiv <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V8i32 = udiv <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V16i32 = udiv <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = udiv i16 undef, 7
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8i16 = udiv <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i16 = udiv <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>

diff --git a/llvm/test/Analysis/CostModel/X86/rem.ll b/llvm/test/Analysis/CostModel/X86/rem.ll
index 7942cda..30dd9a7 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/rem.ll
+++ b/llvm/test/Analysis/CostModel/X86/rem.ll
@@ -450,62 +450,24 @@ define i32 @urem_const() {
 }
 
 define i32 @srem_uniformconst() {
-; SSE2-LABEL: 'srem_uniformconst'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = srem i64 undef, 7
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = srem <2 x i64> undef, <i64 7, i64 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 96 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i16 = srem <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = srem i8 undef, 7
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i8 = srem <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i8 = srem <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V64i8 = srem <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'srem_uniformconst'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = srem i64 undef, 7
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = srem <2 x i64> undef, <i64 7, i64 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 96 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i16 = srem <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = srem i8 undef, 7
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i8 = srem <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i8 = srem <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V64i8 = srem <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSE42-LABEL: 'srem_uniformconst'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = srem i64 undef, 7
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = srem <2 x i64> undef, <i64 7, i64 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i16 = srem <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = srem i8 undef, 7
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i8 = srem <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i8 = srem <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V64i8 = srem <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; SSE-LABEL: 'srem_uniformconst'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = srem i64 undef, 7
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = srem <2 x i64> undef, <i64 7, i64 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i16 = srem <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = srem i8 undef, 7
+; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i8 = srem <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i8 = srem <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V64i8 = srem <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX1-LABEL: 'srem_uniformconst'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = srem i64 undef, 7
@@ -513,9 +475,9 @@ define i32 @srem_uniformconst() {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 76 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -532,9 +494,9 @@ define i32 @srem_uniformconst() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -551,9 +513,9 @@ define i32 @srem_uniformconst() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -570,9 +532,9 @@ define i32 @srem_uniformconst() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -583,53 +545,15 @@ define i32 @srem_uniformconst() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V64i8 = srem <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SLM-LABEL: 'srem_uniformconst'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = srem i64 undef, 7
-; SLM-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = srem <2 x i64> undef, <i64 7, i64 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; SLM-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
-; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i16 = srem <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = srem i8 undef, 7
-; SLM-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i8 = srem <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i8 = srem <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V64i8 = srem <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; GLM-LABEL: 'srem_uniformconst'
-; GLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = srem i64 undef, 7
-; GLM-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = srem <2 x i64> undef, <i64 7, i64 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; GLM-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
-; GLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i16 = srem <32 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I8 = srem i8 undef, 7
-; GLM-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i8 = srem <16 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V32i8 = srem <32 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V64i8 = srem <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
-; GLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; BTVER2-LABEL: 'srem_uniformconst'
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = srem i64 undef, 7
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V2i64 = srem <2 x i64> undef, <i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = srem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = srem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = srem i32 undef, 7
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 76 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V4i32 = srem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V8i32 = srem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V16i32 = srem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = srem i16 undef, 7
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = srem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i16 = srem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -670,9 +594,9 @@ define i32 @urem_uniformconst() {
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = urem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = urem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = urem i32 undef, 7
-; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = urem i16 undef, 7
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = urem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16i16 = urem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -689,9 +613,9 @@ define i32 @urem_uniformconst() {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = urem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = urem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = urem i32 undef, 7
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = urem i16 undef, 7
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = urem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i16 = urem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -708,9 +632,9 @@ define i32 @urem_uniformconst() {
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = urem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = urem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = urem i32 undef, 7
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = urem i16 undef, 7
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = urem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = urem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -727,9 +651,9 @@ define i32 @urem_uniformconst() {
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = urem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = urem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = urem i32 undef, 7
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = urem i16 undef, 7
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = urem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = urem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -746,9 +670,9 @@ define i32 @urem_uniformconst() {
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = urem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = urem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = urem i32 undef, 7
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = urem i16 undef, 7
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = urem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = urem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
@@ -765,9 +689,9 @@ define i32 @urem_uniformconst() {
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V4i64 = urem <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V8i64 = urem <8 x i64> undef, <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I32 = urem i32 undef, 7
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = urem <4 x i32> undef, <i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V8i32 = urem <8 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %V16i32 = urem <16 x i32> undef, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I16 = urem i16 undef, 7
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8i16 = urem <8 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16i16 = urem <16 x i16> undef, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>

diff --git a/llvm/test/Analysis/CostModel/X86/vdiv-cost.ll b/llvm/test/Analysis/CostModel/X86/vdiv-cost.ll
index d87d21c..8552509 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/vdiv-cost.ll
+++ b/llvm/test/Analysis/CostModel/X86/vdiv-cost.ll
@@ -10,7 +10,7 @@
 
 define <4 x i32> @test1(<4 x i32> %a) {
 ; CHECK-LABEL: 'test1'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %div = udiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %div = udiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %div
 ;
   %div = udiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
@@ -19,19 +19,19 @@ define <4 x i32> @test1(<4 x i32> %a) {
 
 define <8 x i32> @test2(<8 x i32> %a) {
 ; SSE-LABEL: 'test2'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
 ;
 ; AVX1-LABEL: 'test2'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
 ;
 ; AVX2-LABEL: 'test2'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
 ;
 ; AVX512-LABEL: 'test2'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
 ;
   %div = udiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7,i32 7, i32 7, i32 7, i32 7>
@@ -108,53 +108,29 @@ define <16 x i8> @test7(<16 x i8> %a) {
 }
 
 define <4 x i32> @test8(<4 x i32> %a) {
-; SSE2-LABEL: 'test8'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %div = sdiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %div
-;
-; SSSE3-LABEL: 'test8'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %div = sdiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %div
-;
-; SSE42-LABEL: 'test8'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %div = sdiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %div
-;
-; AVX-LABEL: 'test8'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %div = sdiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %div
-;
-; AVX512-LABEL: 'test8'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %div = sdiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %div
+; CHECK-LABEL: 'test8'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %div = sdiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %div
 ;
   %div = sdiv <4 x i32> %a, <i32 7, i32 7, i32 7, i32 7>
   ret <4 x i32> %div
 }
 
 define <8 x i32> @test9(<8 x i32> %a) {
-; SSE2-LABEL: 'test9'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
-;
-; SSSE3-LABEL: 'test9'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 38 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
-;
-; SSE42-LABEL: 'test9'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
+; SSE-LABEL: 'test9'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
 ;
 ; AVX1-LABEL: 'test9'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
 ;
 ; AVX2-LABEL: 'test9'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
 ;
 ; AVX512-LABEL: 'test9'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %div
 ;
   %div = sdiv <8 x i32> %a, <i32 7, i32 7, i32 7, i32 7,i32 7, i32 7, i32 7, i32 7>

diff --git a/llvm/test/Transforms/SLPVectorizer/X86/arith-div.ll b/llvm/test/Transforms/SLPVectorizer/X86/arith-div.ll
index 30930ea..fb4ec00 100644 (file)
--- a/llvm/test/Transforms/SLPVectorizer/X86/arith-div.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/arith-div.ll
@@ -24,166 +24,43 @@
 
 define void @sdiv_v16i32_uniformconst() {
 ; SSE-LABEL: @sdiv_v16i32_uniformconst(
-; SSE-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; SSE-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; SSE-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; SSE-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; SSE-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; SSE-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; SSE-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; SSE-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; SSE-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; SSE-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; SSE-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; SSE-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; SSE-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; SSE-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; SSE-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; SSE-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; SSE-NEXT:    [[R0:%.*]] = sdiv i32 [[A0]], 5
-; SSE-NEXT:    [[R1:%.*]] = sdiv i32 [[A1]], 5
-; SSE-NEXT:    [[R2:%.*]] = sdiv i32 [[A2]], 5
-; SSE-NEXT:    [[R3:%.*]] = sdiv i32 [[A3]], 5
-; SSE-NEXT:    [[R4:%.*]] = sdiv i32 [[A4]], 5
-; SSE-NEXT:    [[R5:%.*]] = sdiv i32 [[A5]], 5
-; SSE-NEXT:    [[R6:%.*]] = sdiv i32 [[A6]], 5
-; SSE-NEXT:    [[R7:%.*]] = sdiv i32 [[A7]], 5
-; SSE-NEXT:    [[R8:%.*]] = sdiv i32 [[A8]], 5
-; SSE-NEXT:    [[R9:%.*]] = sdiv i32 [[A9]], 5
-; SSE-NEXT:    [[R10:%.*]] = sdiv i32 [[A10]], 5
-; SSE-NEXT:    [[R11:%.*]] = sdiv i32 [[A11]], 5
-; SSE-NEXT:    [[R12:%.*]] = sdiv i32 [[A12]], 5
-; SSE-NEXT:    [[R13:%.*]] = sdiv i32 [[A13]], 5
-; SSE-NEXT:    [[R14:%.*]] = sdiv i32 [[A14]], 5
-; SSE-NEXT:    [[R15:%.*]] = sdiv i32 [[A15]], 5
-; SSE-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; SSE-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; SSE-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; SSE-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; SSE-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; SSE-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; SSE-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; SSE-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; SSE-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; SSE-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; SSE-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; SSE-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; SSE-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; SSE-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; SSE-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; SSE-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
+; SSE-NEXT:    [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([16 x i32]* @a32 to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP5:%.*]] = sdiv <4 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP6:%.*]] = sdiv <4 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP7:%.*]] = sdiv <4 x i32> [[TMP3]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP8:%.*]] = sdiv <4 x i32> [[TMP4]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    store <4 x i32> [[TMP5]], <4 x i32>* bitcast ([16 x i32]* @c32 to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4) to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP8]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12) to <4 x i32>*), align 4
 ; SSE-NEXT:    ret void
 ;
 ; SLM-LABEL: @sdiv_v16i32_uniformconst(
-; SLM-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; SLM-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; SLM-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; SLM-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; SLM-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; SLM-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; SLM-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; SLM-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; SLM-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; SLM-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; SLM-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; SLM-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; SLM-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; SLM-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; SLM-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; SLM-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; SLM-NEXT:    [[R0:%.*]] = sdiv i32 [[A0]], 5
-; SLM-NEXT:    [[R1:%.*]] = sdiv i32 [[A1]], 5
-; SLM-NEXT:    [[R2:%.*]] = sdiv i32 [[A2]], 5
-; SLM-NEXT:    [[R3:%.*]] = sdiv i32 [[A3]], 5
-; SLM-NEXT:    [[R4:%.*]] = sdiv i32 [[A4]], 5
-; SLM-NEXT:    [[R5:%.*]] = sdiv i32 [[A5]], 5
-; SLM-NEXT:    [[R6:%.*]] = sdiv i32 [[A6]], 5
-; SLM-NEXT:    [[R7:%.*]] = sdiv i32 [[A7]], 5
-; SLM-NEXT:    [[R8:%.*]] = sdiv i32 [[A8]], 5
-; SLM-NEXT:    [[R9:%.*]] = sdiv i32 [[A9]], 5
-; SLM-NEXT:    [[R10:%.*]] = sdiv i32 [[A10]], 5
-; SLM-NEXT:    [[R11:%.*]] = sdiv i32 [[A11]], 5
-; SLM-NEXT:    [[R12:%.*]] = sdiv i32 [[A12]], 5
-; SLM-NEXT:    [[R13:%.*]] = sdiv i32 [[A13]], 5
-; SLM-NEXT:    [[R14:%.*]] = sdiv i32 [[A14]], 5
-; SLM-NEXT:    [[R15:%.*]] = sdiv i32 [[A15]], 5
-; SLM-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; SLM-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; SLM-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; SLM-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; SLM-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; SLM-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; SLM-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; SLM-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; SLM-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; SLM-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; SLM-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; SLM-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; SLM-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; SLM-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; SLM-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; SLM-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
+; SLM-NEXT:    [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([16 x i32]* @a32 to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP5:%.*]] = sdiv <4 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP6:%.*]] = sdiv <4 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP7:%.*]] = sdiv <4 x i32> [[TMP3]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP8:%.*]] = sdiv <4 x i32> [[TMP4]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    store <4 x i32> [[TMP5]], <4 x i32>* bitcast ([16 x i32]* @c32 to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4) to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP8]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12) to <4 x i32>*), align 4
 ; SLM-NEXT:    ret void
 ;
-; AVX1-LABEL: @sdiv_v16i32_uniformconst(
-; AVX1-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; AVX1-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; AVX1-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; AVX1-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; AVX1-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; AVX1-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; AVX1-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; AVX1-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; AVX1-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; AVX1-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; AVX1-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; AVX1-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; AVX1-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; AVX1-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; AVX1-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; AVX1-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; AVX1-NEXT:    [[R0:%.*]] = sdiv i32 [[A0]], 5
-; AVX1-NEXT:    [[R1:%.*]] = sdiv i32 [[A1]], 5
-; AVX1-NEXT:    [[R2:%.*]] = sdiv i32 [[A2]], 5
-; AVX1-NEXT:    [[R3:%.*]] = sdiv i32 [[A3]], 5
-; AVX1-NEXT:    [[R4:%.*]] = sdiv i32 [[A4]], 5
-; AVX1-NEXT:    [[R5:%.*]] = sdiv i32 [[A5]], 5
-; AVX1-NEXT:    [[R6:%.*]] = sdiv i32 [[A6]], 5
-; AVX1-NEXT:    [[R7:%.*]] = sdiv i32 [[A7]], 5
-; AVX1-NEXT:    [[R8:%.*]] = sdiv i32 [[A8]], 5
-; AVX1-NEXT:    [[R9:%.*]] = sdiv i32 [[A9]], 5
-; AVX1-NEXT:    [[R10:%.*]] = sdiv i32 [[A10]], 5
-; AVX1-NEXT:    [[R11:%.*]] = sdiv i32 [[A11]], 5
-; AVX1-NEXT:    [[R12:%.*]] = sdiv i32 [[A12]], 5
-; AVX1-NEXT:    [[R13:%.*]] = sdiv i32 [[A13]], 5
-; AVX1-NEXT:    [[R14:%.*]] = sdiv i32 [[A14]], 5
-; AVX1-NEXT:    [[R15:%.*]] = sdiv i32 [[A15]], 5
-; AVX1-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; AVX1-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; AVX1-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; AVX1-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; AVX1-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; AVX1-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; AVX1-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; AVX1-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; AVX1-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; AVX1-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; AVX1-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; AVX1-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; AVX1-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; AVX1-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; AVX1-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; AVX1-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
-; AVX1-NEXT:    ret void
-;
-; AVX2-LABEL: @sdiv_v16i32_uniformconst(
-; AVX2-NEXT:    [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
-; AVX2-NEXT:    [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT:    [[TMP3:%.*]] = sdiv <8 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
-; AVX2-NEXT:    [[TMP4:%.*]] = sdiv <8 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
-; AVX2-NEXT:    store <8 x i32> [[TMP3]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
-; AVX2-NEXT:    store <8 x i32> [[TMP4]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT:    ret void
+; AVX-LABEL: @sdiv_v16i32_uniformconst(
+; AVX-NEXT:    [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
+; AVX-NEXT:    [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT:    [[TMP3:%.*]] = sdiv <8 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+; AVX-NEXT:    [[TMP4:%.*]] = sdiv <8 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+; AVX-NEXT:    store <8 x i32> [[TMP3]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
+; AVX-NEXT:    store <8 x i32> [[TMP4]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT:    ret void
 ;
 ; AVX512-LABEL: @sdiv_v16i32_uniformconst(
 ; AVX512-NEXT:    [[TMP1:%.*]] = load <16 x i32>, <16 x i32>* bitcast ([16 x i32]* @a32 to <16 x i32>*), align 4
@@ -244,166 +121,43 @@ define void @sdiv_v16i32_uniformconst() {
 
 define void @srem_v16i32_uniformconst() {
 ; SSE-LABEL: @srem_v16i32_uniformconst(
-; SSE-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; SSE-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; SSE-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; SSE-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; SSE-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; SSE-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; SSE-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; SSE-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; SSE-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; SSE-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; SSE-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; SSE-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; SSE-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; SSE-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; SSE-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; SSE-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; SSE-NEXT:    [[R0:%.*]] = srem i32 [[A0]], 5
-; SSE-NEXT:    [[R1:%.*]] = srem i32 [[A1]], 5
-; SSE-NEXT:    [[R2:%.*]] = srem i32 [[A2]], 5
-; SSE-NEXT:    [[R3:%.*]] = srem i32 [[A3]], 5
-; SSE-NEXT:    [[R4:%.*]] = srem i32 [[A4]], 5
-; SSE-NEXT:    [[R5:%.*]] = srem i32 [[A5]], 5
-; SSE-NEXT:    [[R6:%.*]] = srem i32 [[A6]], 5
-; SSE-NEXT:    [[R7:%.*]] = srem i32 [[A7]], 5
-; SSE-NEXT:    [[R8:%.*]] = srem i32 [[A8]], 5
-; SSE-NEXT:    [[R9:%.*]] = srem i32 [[A9]], 5
-; SSE-NEXT:    [[R10:%.*]] = srem i32 [[A10]], 5
-; SSE-NEXT:    [[R11:%.*]] = srem i32 [[A11]], 5
-; SSE-NEXT:    [[R12:%.*]] = srem i32 [[A12]], 5
-; SSE-NEXT:    [[R13:%.*]] = srem i32 [[A13]], 5
-; SSE-NEXT:    [[R14:%.*]] = srem i32 [[A14]], 5
-; SSE-NEXT:    [[R15:%.*]] = srem i32 [[A15]], 5
-; SSE-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; SSE-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; SSE-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; SSE-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; SSE-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; SSE-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; SSE-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; SSE-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; SSE-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; SSE-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; SSE-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; SSE-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; SSE-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; SSE-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; SSE-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; SSE-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
+; SSE-NEXT:    [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([16 x i32]* @a32 to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP5:%.*]] = srem <4 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP6:%.*]] = srem <4 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP7:%.*]] = srem <4 x i32> [[TMP3]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP8:%.*]] = srem <4 x i32> [[TMP4]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    store <4 x i32> [[TMP5]], <4 x i32>* bitcast ([16 x i32]* @c32 to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4) to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP8]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12) to <4 x i32>*), align 4
 ; SSE-NEXT:    ret void
 ;
 ; SLM-LABEL: @srem_v16i32_uniformconst(
-; SLM-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; SLM-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; SLM-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; SLM-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; SLM-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; SLM-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; SLM-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; SLM-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; SLM-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; SLM-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; SLM-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; SLM-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; SLM-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; SLM-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; SLM-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; SLM-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; SLM-NEXT:    [[R0:%.*]] = srem i32 [[A0]], 5
-; SLM-NEXT:    [[R1:%.*]] = srem i32 [[A1]], 5
-; SLM-NEXT:    [[R2:%.*]] = srem i32 [[A2]], 5
-; SLM-NEXT:    [[R3:%.*]] = srem i32 [[A3]], 5
-; SLM-NEXT:    [[R4:%.*]] = srem i32 [[A4]], 5
-; SLM-NEXT:    [[R5:%.*]] = srem i32 [[A5]], 5
-; SLM-NEXT:    [[R6:%.*]] = srem i32 [[A6]], 5
-; SLM-NEXT:    [[R7:%.*]] = srem i32 [[A7]], 5
-; SLM-NEXT:    [[R8:%.*]] = srem i32 [[A8]], 5
-; SLM-NEXT:    [[R9:%.*]] = srem i32 [[A9]], 5
-; SLM-NEXT:    [[R10:%.*]] = srem i32 [[A10]], 5
-; SLM-NEXT:    [[R11:%.*]] = srem i32 [[A11]], 5
-; SLM-NEXT:    [[R12:%.*]] = srem i32 [[A12]], 5
-; SLM-NEXT:    [[R13:%.*]] = srem i32 [[A13]], 5
-; SLM-NEXT:    [[R14:%.*]] = srem i32 [[A14]], 5
-; SLM-NEXT:    [[R15:%.*]] = srem i32 [[A15]], 5
-; SLM-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; SLM-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; SLM-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; SLM-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; SLM-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; SLM-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; SLM-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; SLM-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; SLM-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; SLM-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; SLM-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; SLM-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; SLM-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; SLM-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; SLM-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; SLM-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
+; SLM-NEXT:    [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([16 x i32]* @a32 to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP5:%.*]] = srem <4 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP6:%.*]] = srem <4 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP7:%.*]] = srem <4 x i32> [[TMP3]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP8:%.*]] = srem <4 x i32> [[TMP4]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    store <4 x i32> [[TMP5]], <4 x i32>* bitcast ([16 x i32]* @c32 to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4) to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP8]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12) to <4 x i32>*), align 4
 ; SLM-NEXT:    ret void
 ;
-; AVX1-LABEL: @srem_v16i32_uniformconst(
-; AVX1-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; AVX1-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; AVX1-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; AVX1-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; AVX1-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; AVX1-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; AVX1-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; AVX1-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; AVX1-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; AVX1-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; AVX1-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; AVX1-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; AVX1-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; AVX1-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; AVX1-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; AVX1-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; AVX1-NEXT:    [[R0:%.*]] = srem i32 [[A0]], 5
-; AVX1-NEXT:    [[R1:%.*]] = srem i32 [[A1]], 5
-; AVX1-NEXT:    [[R2:%.*]] = srem i32 [[A2]], 5
-; AVX1-NEXT:    [[R3:%.*]] = srem i32 [[A3]], 5
-; AVX1-NEXT:    [[R4:%.*]] = srem i32 [[A4]], 5
-; AVX1-NEXT:    [[R5:%.*]] = srem i32 [[A5]], 5
-; AVX1-NEXT:    [[R6:%.*]] = srem i32 [[A6]], 5
-; AVX1-NEXT:    [[R7:%.*]] = srem i32 [[A7]], 5
-; AVX1-NEXT:    [[R8:%.*]] = srem i32 [[A8]], 5
-; AVX1-NEXT:    [[R9:%.*]] = srem i32 [[A9]], 5
-; AVX1-NEXT:    [[R10:%.*]] = srem i32 [[A10]], 5
-; AVX1-NEXT:    [[R11:%.*]] = srem i32 [[A11]], 5
-; AVX1-NEXT:    [[R12:%.*]] = srem i32 [[A12]], 5
-; AVX1-NEXT:    [[R13:%.*]] = srem i32 [[A13]], 5
-; AVX1-NEXT:    [[R14:%.*]] = srem i32 [[A14]], 5
-; AVX1-NEXT:    [[R15:%.*]] = srem i32 [[A15]], 5
-; AVX1-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; AVX1-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; AVX1-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; AVX1-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; AVX1-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; AVX1-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; AVX1-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; AVX1-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; AVX1-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; AVX1-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; AVX1-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; AVX1-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; AVX1-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; AVX1-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; AVX1-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; AVX1-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
-; AVX1-NEXT:    ret void
-;
-; AVX2-LABEL: @srem_v16i32_uniformconst(
-; AVX2-NEXT:    [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
-; AVX2-NEXT:    [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT:    [[TMP3:%.*]] = srem <8 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
-; AVX2-NEXT:    [[TMP4:%.*]] = srem <8 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
-; AVX2-NEXT:    store <8 x i32> [[TMP3]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
-; AVX2-NEXT:    store <8 x i32> [[TMP4]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT:    ret void
+; AVX-LABEL: @srem_v16i32_uniformconst(
+; AVX-NEXT:    [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
+; AVX-NEXT:    [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT:    [[TMP3:%.*]] = srem <8 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+; AVX-NEXT:    [[TMP4:%.*]] = srem <8 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+; AVX-NEXT:    store <8 x i32> [[TMP3]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
+; AVX-NEXT:    store <8 x i32> [[TMP4]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT:    ret void
 ;
 ; AVX512-LABEL: @srem_v16i32_uniformconst(
 ; AVX512-NEXT:    [[TMP1:%.*]] = load <16 x i32>, <16 x i32>* bitcast ([16 x i32]* @a32 to <16 x i32>*), align 4
@@ -464,166 +218,43 @@ define void @srem_v16i32_uniformconst() {
 
 define void @udiv_v16i32_uniformconst() {
 ; SSE-LABEL: @udiv_v16i32_uniformconst(
-; SSE-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; SSE-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; SSE-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; SSE-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; SSE-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; SSE-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; SSE-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; SSE-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; SSE-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; SSE-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; SSE-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; SSE-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; SSE-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; SSE-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; SSE-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; SSE-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; SSE-NEXT:    [[R0:%.*]] = udiv i32 [[A0]], 5
-; SSE-NEXT:    [[R1:%.*]] = udiv i32 [[A1]], 5
-; SSE-NEXT:    [[R2:%.*]] = udiv i32 [[A2]], 5
-; SSE-NEXT:    [[R3:%.*]] = udiv i32 [[A3]], 5
-; SSE-NEXT:    [[R4:%.*]] = udiv i32 [[A4]], 5
-; SSE-NEXT:    [[R5:%.*]] = udiv i32 [[A5]], 5
-; SSE-NEXT:    [[R6:%.*]] = udiv i32 [[A6]], 5
-; SSE-NEXT:    [[R7:%.*]] = udiv i32 [[A7]], 5
-; SSE-NEXT:    [[R8:%.*]] = udiv i32 [[A8]], 5
-; SSE-NEXT:    [[R9:%.*]] = udiv i32 [[A9]], 5
-; SSE-NEXT:    [[R10:%.*]] = udiv i32 [[A10]], 5
-; SSE-NEXT:    [[R11:%.*]] = udiv i32 [[A11]], 5
-; SSE-NEXT:    [[R12:%.*]] = udiv i32 [[A12]], 5
-; SSE-NEXT:    [[R13:%.*]] = udiv i32 [[A13]], 5
-; SSE-NEXT:    [[R14:%.*]] = udiv i32 [[A14]], 5
-; SSE-NEXT:    [[R15:%.*]] = udiv i32 [[A15]], 5
-; SSE-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; SSE-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; SSE-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; SSE-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; SSE-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; SSE-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; SSE-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; SSE-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; SSE-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; SSE-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; SSE-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; SSE-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; SSE-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; SSE-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; SSE-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; SSE-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
+; SSE-NEXT:    [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([16 x i32]* @a32 to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP5:%.*]] = udiv <4 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP6:%.*]] = udiv <4 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP7:%.*]] = udiv <4 x i32> [[TMP3]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP8:%.*]] = udiv <4 x i32> [[TMP4]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    store <4 x i32> [[TMP5]], <4 x i32>* bitcast ([16 x i32]* @c32 to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4) to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP8]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12) to <4 x i32>*), align 4
 ; SSE-NEXT:    ret void
 ;
 ; SLM-LABEL: @udiv_v16i32_uniformconst(
-; SLM-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; SLM-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; SLM-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; SLM-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; SLM-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; SLM-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; SLM-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; SLM-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; SLM-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; SLM-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; SLM-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; SLM-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; SLM-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; SLM-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; SLM-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; SLM-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; SLM-NEXT:    [[R0:%.*]] = udiv i32 [[A0]], 5
-; SLM-NEXT:    [[R1:%.*]] = udiv i32 [[A1]], 5
-; SLM-NEXT:    [[R2:%.*]] = udiv i32 [[A2]], 5
-; SLM-NEXT:    [[R3:%.*]] = udiv i32 [[A3]], 5
-; SLM-NEXT:    [[R4:%.*]] = udiv i32 [[A4]], 5
-; SLM-NEXT:    [[R5:%.*]] = udiv i32 [[A5]], 5
-; SLM-NEXT:    [[R6:%.*]] = udiv i32 [[A6]], 5
-; SLM-NEXT:    [[R7:%.*]] = udiv i32 [[A7]], 5
-; SLM-NEXT:    [[R8:%.*]] = udiv i32 [[A8]], 5
-; SLM-NEXT:    [[R9:%.*]] = udiv i32 [[A9]], 5
-; SLM-NEXT:    [[R10:%.*]] = udiv i32 [[A10]], 5
-; SLM-NEXT:    [[R11:%.*]] = udiv i32 [[A11]], 5
-; SLM-NEXT:    [[R12:%.*]] = udiv i32 [[A12]], 5
-; SLM-NEXT:    [[R13:%.*]] = udiv i32 [[A13]], 5
-; SLM-NEXT:    [[R14:%.*]] = udiv i32 [[A14]], 5
-; SLM-NEXT:    [[R15:%.*]] = udiv i32 [[A15]], 5
-; SLM-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; SLM-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; SLM-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; SLM-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; SLM-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; SLM-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; SLM-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; SLM-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; SLM-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; SLM-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; SLM-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; SLM-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; SLM-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; SLM-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; SLM-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; SLM-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
+; SLM-NEXT:    [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([16 x i32]* @a32 to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP5:%.*]] = udiv <4 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP6:%.*]] = udiv <4 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP7:%.*]] = udiv <4 x i32> [[TMP3]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP8:%.*]] = udiv <4 x i32> [[TMP4]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    store <4 x i32> [[TMP5]], <4 x i32>* bitcast ([16 x i32]* @c32 to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4) to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP8]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12) to <4 x i32>*), align 4
 ; SLM-NEXT:    ret void
 ;
-; AVX1-LABEL: @udiv_v16i32_uniformconst(
-; AVX1-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; AVX1-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; AVX1-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; AVX1-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; AVX1-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; AVX1-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; AVX1-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; AVX1-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; AVX1-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; AVX1-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; AVX1-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; AVX1-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; AVX1-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; AVX1-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; AVX1-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; AVX1-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; AVX1-NEXT:    [[R0:%.*]] = udiv i32 [[A0]], 5
-; AVX1-NEXT:    [[R1:%.*]] = udiv i32 [[A1]], 5
-; AVX1-NEXT:    [[R2:%.*]] = udiv i32 [[A2]], 5
-; AVX1-NEXT:    [[R3:%.*]] = udiv i32 [[A3]], 5
-; AVX1-NEXT:    [[R4:%.*]] = udiv i32 [[A4]], 5
-; AVX1-NEXT:    [[R5:%.*]] = udiv i32 [[A5]], 5
-; AVX1-NEXT:    [[R6:%.*]] = udiv i32 [[A6]], 5
-; AVX1-NEXT:    [[R7:%.*]] = udiv i32 [[A7]], 5
-; AVX1-NEXT:    [[R8:%.*]] = udiv i32 [[A8]], 5
-; AVX1-NEXT:    [[R9:%.*]] = udiv i32 [[A9]], 5
-; AVX1-NEXT:    [[R10:%.*]] = udiv i32 [[A10]], 5
-; AVX1-NEXT:    [[R11:%.*]] = udiv i32 [[A11]], 5
-; AVX1-NEXT:    [[R12:%.*]] = udiv i32 [[A12]], 5
-; AVX1-NEXT:    [[R13:%.*]] = udiv i32 [[A13]], 5
-; AVX1-NEXT:    [[R14:%.*]] = udiv i32 [[A14]], 5
-; AVX1-NEXT:    [[R15:%.*]] = udiv i32 [[A15]], 5
-; AVX1-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; AVX1-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; AVX1-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; AVX1-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; AVX1-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; AVX1-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; AVX1-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; AVX1-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; AVX1-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; AVX1-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; AVX1-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; AVX1-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; AVX1-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; AVX1-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; AVX1-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; AVX1-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
-; AVX1-NEXT:    ret void
-;
-; AVX2-LABEL: @udiv_v16i32_uniformconst(
-; AVX2-NEXT:    [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
-; AVX2-NEXT:    [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT:    [[TMP3:%.*]] = udiv <8 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
-; AVX2-NEXT:    [[TMP4:%.*]] = udiv <8 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
-; AVX2-NEXT:    store <8 x i32> [[TMP3]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
-; AVX2-NEXT:    store <8 x i32> [[TMP4]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT:    ret void
+; AVX-LABEL: @udiv_v16i32_uniformconst(
+; AVX-NEXT:    [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
+; AVX-NEXT:    [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT:    [[TMP3:%.*]] = udiv <8 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+; AVX-NEXT:    [[TMP4:%.*]] = udiv <8 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+; AVX-NEXT:    store <8 x i32> [[TMP3]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
+; AVX-NEXT:    store <8 x i32> [[TMP4]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT:    ret void
 ;
 ; AVX512-LABEL: @udiv_v16i32_uniformconst(
 ; AVX512-NEXT:    [[TMP1:%.*]] = load <16 x i32>, <16 x i32>* bitcast ([16 x i32]* @a32 to <16 x i32>*), align 4
@@ -684,166 +315,43 @@ define void @udiv_v16i32_uniformconst() {
 
 define void @urem_v16i32_uniformconst() {
 ; SSE-LABEL: @urem_v16i32_uniformconst(
-; SSE-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; SSE-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; SSE-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; SSE-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; SSE-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; SSE-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; SSE-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; SSE-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; SSE-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; SSE-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; SSE-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; SSE-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; SSE-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; SSE-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; SSE-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; SSE-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; SSE-NEXT:    [[R0:%.*]] = urem i32 [[A0]], 5
-; SSE-NEXT:    [[R1:%.*]] = urem i32 [[A1]], 5
-; SSE-NEXT:    [[R2:%.*]] = urem i32 [[A2]], 5
-; SSE-NEXT:    [[R3:%.*]] = urem i32 [[A3]], 5
-; SSE-NEXT:    [[R4:%.*]] = urem i32 [[A4]], 5
-; SSE-NEXT:    [[R5:%.*]] = urem i32 [[A5]], 5
-; SSE-NEXT:    [[R6:%.*]] = urem i32 [[A6]], 5
-; SSE-NEXT:    [[R7:%.*]] = urem i32 [[A7]], 5
-; SSE-NEXT:    [[R8:%.*]] = urem i32 [[A8]], 5
-; SSE-NEXT:    [[R9:%.*]] = urem i32 [[A9]], 5
-; SSE-NEXT:    [[R10:%.*]] = urem i32 [[A10]], 5
-; SSE-NEXT:    [[R11:%.*]] = urem i32 [[A11]], 5
-; SSE-NEXT:    [[R12:%.*]] = urem i32 [[A12]], 5
-; SSE-NEXT:    [[R13:%.*]] = urem i32 [[A13]], 5
-; SSE-NEXT:    [[R14:%.*]] = urem i32 [[A14]], 5
-; SSE-NEXT:    [[R15:%.*]] = urem i32 [[A15]], 5
-; SSE-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; SSE-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; SSE-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; SSE-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; SSE-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; SSE-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; SSE-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; SSE-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; SSE-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; SSE-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; SSE-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; SSE-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; SSE-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; SSE-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; SSE-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; SSE-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
+; SSE-NEXT:    [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([16 x i32]* @a32 to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12) to <4 x i32>*), align 4
+; SSE-NEXT:    [[TMP5:%.*]] = urem <4 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP6:%.*]] = urem <4 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP7:%.*]] = urem <4 x i32> [[TMP3]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    [[TMP8:%.*]] = urem <4 x i32> [[TMP4]], <i32 5, i32 5, i32 5, i32 5>
+; SSE-NEXT:    store <4 x i32> [[TMP5]], <4 x i32>* bitcast ([16 x i32]* @c32 to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4) to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <4 x i32>*), align 4
+; SSE-NEXT:    store <4 x i32> [[TMP8]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12) to <4 x i32>*), align 4
 ; SSE-NEXT:    ret void
 ;
 ; SLM-LABEL: @urem_v16i32_uniformconst(
-; SLM-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; SLM-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; SLM-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; SLM-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; SLM-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; SLM-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; SLM-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; SLM-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; SLM-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; SLM-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; SLM-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; SLM-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; SLM-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; SLM-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; SLM-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; SLM-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; SLM-NEXT:    [[R0:%.*]] = urem i32 [[A0]], 5
-; SLM-NEXT:    [[R1:%.*]] = urem i32 [[A1]], 5
-; SLM-NEXT:    [[R2:%.*]] = urem i32 [[A2]], 5
-; SLM-NEXT:    [[R3:%.*]] = urem i32 [[A3]], 5
-; SLM-NEXT:    [[R4:%.*]] = urem i32 [[A4]], 5
-; SLM-NEXT:    [[R5:%.*]] = urem i32 [[A5]], 5
-; SLM-NEXT:    [[R6:%.*]] = urem i32 [[A6]], 5
-; SLM-NEXT:    [[R7:%.*]] = urem i32 [[A7]], 5
-; SLM-NEXT:    [[R8:%.*]] = urem i32 [[A8]], 5
-; SLM-NEXT:    [[R9:%.*]] = urem i32 [[A9]], 5
-; SLM-NEXT:    [[R10:%.*]] = urem i32 [[A10]], 5
-; SLM-NEXT:    [[R11:%.*]] = urem i32 [[A11]], 5
-; SLM-NEXT:    [[R12:%.*]] = urem i32 [[A12]], 5
-; SLM-NEXT:    [[R13:%.*]] = urem i32 [[A13]], 5
-; SLM-NEXT:    [[R14:%.*]] = urem i32 [[A14]], 5
-; SLM-NEXT:    [[R15:%.*]] = urem i32 [[A15]], 5
-; SLM-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; SLM-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; SLM-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; SLM-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; SLM-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; SLM-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; SLM-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; SLM-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; SLM-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; SLM-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; SLM-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; SLM-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; SLM-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; SLM-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; SLM-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; SLM-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
+; SLM-NEXT:    [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([16 x i32]* @a32 to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12) to <4 x i32>*), align 4
+; SLM-NEXT:    [[TMP5:%.*]] = urem <4 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP6:%.*]] = urem <4 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP7:%.*]] = urem <4 x i32> [[TMP3]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    [[TMP8:%.*]] = urem <4 x i32> [[TMP4]], <i32 5, i32 5, i32 5, i32 5>
+; SLM-NEXT:    store <4 x i32> [[TMP5]], <4 x i32>* bitcast ([16 x i32]* @c32 to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4) to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <4 x i32>*), align 4
+; SLM-NEXT:    store <4 x i32> [[TMP8]], <4 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12) to <4 x i32>*), align 4
 ; SLM-NEXT:    ret void
 ;
-; AVX1-LABEL: @urem_v16i32_uniformconst(
-; AVX1-NEXT:    [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; AVX1-NEXT:    [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; AVX1-NEXT:    [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; AVX1-NEXT:    [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; AVX1-NEXT:    [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; AVX1-NEXT:    [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; AVX1-NEXT:    [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; AVX1-NEXT:    [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; AVX1-NEXT:    [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; AVX1-NEXT:    [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; AVX1-NEXT:    [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; AVX1-NEXT:    [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; AVX1-NEXT:    [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; AVX1-NEXT:    [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; AVX1-NEXT:    [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; AVX1-NEXT:    [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; AVX1-NEXT:    [[R0:%.*]] = urem i32 [[A0]], 5
-; AVX1-NEXT:    [[R1:%.*]] = urem i32 [[A1]], 5
-; AVX1-NEXT:    [[R2:%.*]] = urem i32 [[A2]], 5
-; AVX1-NEXT:    [[R3:%.*]] = urem i32 [[A3]], 5
-; AVX1-NEXT:    [[R4:%.*]] = urem i32 [[A4]], 5
-; AVX1-NEXT:    [[R5:%.*]] = urem i32 [[A5]], 5
-; AVX1-NEXT:    [[R6:%.*]] = urem i32 [[A6]], 5
-; AVX1-NEXT:    [[R7:%.*]] = urem i32 [[A7]], 5
-; AVX1-NEXT:    [[R8:%.*]] = urem i32 [[A8]], 5
-; AVX1-NEXT:    [[R9:%.*]] = urem i32 [[A9]], 5
-; AVX1-NEXT:    [[R10:%.*]] = urem i32 [[A10]], 5
-; AVX1-NEXT:    [[R11:%.*]] = urem i32 [[A11]], 5
-; AVX1-NEXT:    [[R12:%.*]] = urem i32 [[A12]], 5
-; AVX1-NEXT:    [[R13:%.*]] = urem i32 [[A13]], 5
-; AVX1-NEXT:    [[R14:%.*]] = urem i32 [[A14]], 5
-; AVX1-NEXT:    [[R15:%.*]] = urem i32 [[A15]], 5
-; AVX1-NEXT:    store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; AVX1-NEXT:    store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; AVX1-NEXT:    store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; AVX1-NEXT:    store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; AVX1-NEXT:    store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; AVX1-NEXT:    store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; AVX1-NEXT:    store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; AVX1-NEXT:    store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; AVX1-NEXT:    store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; AVX1-NEXT:    store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; AVX1-NEXT:    store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; AVX1-NEXT:    store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; AVX1-NEXT:    store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; AVX1-NEXT:    store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; AVX1-NEXT:    store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; AVX1-NEXT:    store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
-; AVX1-NEXT:    ret void
-;
-; AVX2-LABEL: @urem_v16i32_uniformconst(
-; AVX2-NEXT:    [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
-; AVX2-NEXT:    [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT:    [[TMP3:%.*]] = urem <8 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
-; AVX2-NEXT:    [[TMP4:%.*]] = urem <8 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
-; AVX2-NEXT:    store <8 x i32> [[TMP3]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
-; AVX2-NEXT:    store <8 x i32> [[TMP4]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT:    ret void
+; AVX-LABEL: @urem_v16i32_uniformconst(
+; AVX-NEXT:    [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
+; AVX-NEXT:    [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT:    [[TMP3:%.*]] = urem <8 x i32> [[TMP1]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+; AVX-NEXT:    [[TMP4:%.*]] = urem <8 x i32> [[TMP2]], <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+; AVX-NEXT:    store <8 x i32> [[TMP3]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
+; AVX-NEXT:    store <8 x i32> [[TMP4]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT:    ret void
 ;
 ; AVX512-LABEL: @urem_v16i32_uniformconst(
 ; AVX512-NEXT:    [[TMP1:%.*]] = load <16 x i32>, <16 x i32>* bitcast ([16 x i32]* @a32 to <16 x i32>*), align 4