[X86][CostModel] Correct the costs for truncate to a mask register with avx512

I've modified isTruncateFree to get an accurate cost for types that need to be split. I'm planning to look into fixing it for all vectors, but need more cost cleanups first.

Differential Revision: https://reviews.llvm.org/D78973

diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index 9f0f3cd..cf70f7e 100644 (file)
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -30611,6 +30611,11 @@ bool X86TargetLowering::isLegalStoreImmediate(int64_t Imm) const {
 bool X86TargetLowering::isTruncateFree(EVT VT1, EVT VT2) const {
   if (!VT1.isInteger() || !VT2.isInteger())
     return false;
+  // Truncate to mask registers aren't free.
+  // TODO: No vector truncates are free.
+  if (Subtarget.hasAVX512() && VT2.isVector() &&
+      VT2.getVectorElementType() == MVT::i1)
+    return false;
   unsigned NumBits1 = VT1.getSizeInBits();
   unsigned NumBits2 = VT2.getSizeInBits();
   return NumBits1 > NumBits2;

diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index c5148f7..21c9b75 100644 (file)
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -1396,6 +1396,17 @@ int X86TTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src,
 
     { ISD::TRUNCATE,    MVT::v32i8,  MVT::v32i16, 2 },
     { ISD::TRUNCATE,    MVT::v16i8,  MVT::v16i16, 2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v2i1,   MVT::v2i8,   2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v2i1,   MVT::v2i16,  2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v4i1,   MVT::v4i8,   2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v4i1,   MVT::v4i16,  2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v8i1,   MVT::v8i8,   2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v8i1,   MVT::v8i16,  2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v16i1,  MVT::v16i8,  2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v16i1,  MVT::v16i16, 2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v32i1,  MVT::v32i8,  2 }, // widen to zmm
+    { ISD::TRUNCATE,    MVT::v32i1,  MVT::v32i16, 2 },
+    { ISD::TRUNCATE,    MVT::v64i1,  MVT::v64i8,  2 },
   };
 
   static const TypeConversionCostTblEntry AVX512DQConversionTbl[] = {
@@ -1420,6 +1431,20 @@ int X86TTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src,
     { ISD::FP_EXTEND, MVT::v8f64,   MVT::v16f32, 3 },
     { ISD::FP_ROUND,  MVT::v8f32,   MVT::v8f64,  1 },
 
+    { ISD::TRUNCATE,  MVT::v2i1,    MVT::v2i8,   3 }, // sext+vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v4i1,    MVT::v4i8,   3 }, // sext+vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v8i1,    MVT::v8i8,   3 }, // sext+vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v16i1,   MVT::v16i8,  3 }, // sext+vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v2i1,    MVT::v2i16,  3 }, // sext+vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v4i1,    MVT::v4i16,  3 }, // sext+vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v8i1,    MVT::v8i16,  3 }, // sext+vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v16i1,   MVT::v16i16, 3 }, // sext+vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v4i1,    MVT::v4i32,  2 }, // zmm vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v8i1,    MVT::v8i32,  2 }, // zmm vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v16i1,   MVT::v16i32, 2 }, // vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v2i1,    MVT::v2i64,  2 }, // zmm vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v4i1,    MVT::v4i64,  2 }, // zmm vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v8i1,    MVT::v8i64,  2 }, // vpsllq+vptestmq
     { ISD::TRUNCATE,  MVT::v16i8,   MVT::v16i32, 2 },
     { ISD::TRUNCATE,  MVT::v16i16,  MVT::v16i32, 2 },
     { ISD::TRUNCATE,  MVT::v8i8,    MVT::v8i64,  2 },
@@ -1539,6 +1564,15 @@ int X86TTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src,
     { ISD::ZERO_EXTEND, MVT::v32i8,  MVT::v32i1, 2 },
 
     { ISD::TRUNCATE,    MVT::v16i8,  MVT::v16i16, 2 },
+    { ISD::TRUNCATE,    MVT::v2i1,   MVT::v2i8,   2 }, // vpsllw+vptestmb
+    { ISD::TRUNCATE,    MVT::v2i1,   MVT::v2i16,  2 }, // vpsllw+vptestmw
+    { ISD::TRUNCATE,    MVT::v4i1,   MVT::v4i8,   2 }, // vpsllw+vptestmb
+    { ISD::TRUNCATE,    MVT::v4i1,   MVT::v4i16,  2 }, // vpsllw+vptestmw
+    { ISD::TRUNCATE,    MVT::v8i1,   MVT::v8i8,   2 }, // vpsllw+vptestmb
+    { ISD::TRUNCATE,    MVT::v8i1,   MVT::v8i16,  2 }, // vpsllw+vptestmw
+    { ISD::TRUNCATE,    MVT::v16i1,  MVT::v16i8,  2 }, // vpsllw+vptestmb
+    { ISD::TRUNCATE,    MVT::v16i1,  MVT::v16i16, 2 }, // vpsllw+vptestmw
+    { ISD::TRUNCATE,    MVT::v32i1,  MVT::v32i8,  2 }, // vpsllw+vptestmb
   };
 
   static const TypeConversionCostTblEntry AVX512DQVLConversionTbl[] = {
@@ -1564,6 +1598,19 @@ int X86TTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src,
   };
 
   static const TypeConversionCostTblEntry AVX512VLConversionTbl[] = {
+    { ISD::TRUNCATE,  MVT::v2i1,    MVT::v2i8,   3 }, // sext+vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v4i1,    MVT::v4i8,   3 }, // sext+vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v8i1,    MVT::v8i8,   3 }, // sext+vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v16i1,   MVT::v16i8,  8 }, // split+2*v8i8
+    { ISD::TRUNCATE,  MVT::v2i1,    MVT::v2i16,  3 }, // sext+vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v4i1,    MVT::v4i16,  3 }, // sext+vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v8i1,    MVT::v8i16,  3 }, // sext+vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v16i1,   MVT::v16i16, 8 }, // split+2*v8i16
+    { ISD::TRUNCATE,  MVT::v4i1,    MVT::v4i32,  2 }, // vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v8i1,    MVT::v8i32,  2 }, // vpslld+vptestmd
+    { ISD::TRUNCATE,  MVT::v2i1,    MVT::v2i64,  2 }, // vpsllq+vptestmq
+    { ISD::TRUNCATE,  MVT::v4i1,    MVT::v4i64,  2 }, // vpsllq+vptestmq
+
     // sign extend is vpcmpeq+maskedmove+vpmovdw+vpacksswb
     // zero extend is vpcmpeq+maskedmove+vpmovdw+vpsrlw+vpackuswb
     { ISD::SIGN_EXTEND, MVT::v2i8,   MVT::v2i1,   5 },

diff --git a/llvm/test/Analysis/CostModel/X86/cast.ll b/llvm/test/Analysis/CostModel/X86/cast.ll
index cd2094a..8d38bc6 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/cast.ll
+++ b/llvm/test/Analysis/CostModel/X86/cast.ll
@@ -44,10 +44,10 @@ define i32 @add(i32 %arg) {
 ; AVX512-LABEL: 'add'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %A = zext <4 x i1> undef to <4 x i32>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %B = sext <4 x i1> undef to <4 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %C = trunc <4 x i32> undef to <4 x i1>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %C = trunc <4 x i32> undef to <4 x i1>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %D = zext <8 x i1> undef to <8 x i32>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %E = sext <8 x i1> undef to <8 x i32>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %F = trunc <8 x i32> undef to <8 x i1>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %F = trunc <8 x i32> undef to <8 x i1>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %G = zext i1 undef to i32
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %H = trunc i32 undef to i1
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef

diff --git a/llvm/test/Analysis/CostModel/X86/min-legal-vector-width.ll b/llvm/test/Analysis/CostModel/X86/min-legal-vector-width.ll
index e44ad04..e8e1017 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/min-legal-vector-width.ll
+++ b/llvm/test/Analysis/CostModel/X86/min-legal-vector-width.ll
@@ -549,86 +549,113 @@ define i32 @trunc_vXi1() "min-legal-vector-width"="256" {
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512VL256-LABEL: 'trunc_vXi1'
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
-; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 35 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
+; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 35 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
 ; AVX512VL256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; VEC512-LABEL: 'trunc_vXi1'
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
-; VEC512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX512VL512-LABEL: 'trunc_vXi1'
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
+; AVX512VL512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SKX256-LABEL: 'trunc_vXi1'
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
-; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 47 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
+; SKX256-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
 ; SKX256-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+; SKX512-LABEL: 'trunc_vXi1'
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
+; SKX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
   %V2i64 = trunc <2 x i64> undef to <2 x i1>
   %V4i64 = trunc <4 x i64> undef to <4 x i1>
   %V8i64 = trunc <8 x i64> undef to <8 x i1>

diff --git a/llvm/test/Analysis/CostModel/X86/trunc.ll b/llvm/test/Analysis/CostModel/X86/trunc.ll
index 85ed9d0..6523e3e 100644 (file)
--- a/llvm/test/Analysis/CostModel/X86/trunc.ll
+++ b/llvm/test/Analysis/CostModel/X86/trunc.ll
@@ -408,32 +408,59 @@ define i32 @trunc_vXi1() {
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; AVX512-LABEL: 'trunc_vXi1'
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX512F-LABEL: 'trunc_vXi1'
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX512BW-LABEL: 'trunc_vXi1'
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i64 = trunc <2 x i64> undef to <2 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i64 = trunc <4 x i64> undef to <4 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i64 = trunc <8 x i64> undef to <8 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16i64 = trunc <16 x i64> undef to <16 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V32i64 = trunc <32 x i64> undef to <32 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V64i64 = trunc <64 x i64> undef to <64 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2i32 = trunc <2 x i32> undef to <2 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i32 = trunc <4 x i32> undef to <4 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i32 = trunc <8 x i32> undef to <8 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i32 = trunc <16 x i32> undef to <16 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V32i32 = trunc <32 x i32> undef to <32 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V64i32 = trunc <64 x i32> undef to <64 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i16 = trunc <2 x i16> undef to <2 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i16 = trunc <4 x i16> undef to <4 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i16 = trunc <8 x i16> undef to <8 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i16 = trunc <16 x i16> undef to <16 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i16 = trunc <32 x i16> undef to <32 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64i16 = trunc <64 x i16> undef to <64 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2i8 = trunc <2 x i8> undef to <2 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4i8 = trunc <4 x i8> undef to <4 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8i8 = trunc <8 x i8> undef to <8 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16i8 = trunc <16 x i8> undef to <16 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32i8 = trunc <32 x i8> undef to <32 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64i8 = trunc <64 x i8> undef to <64 x i1>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
   %V2i64 = trunc <2 x i64> undef to <2 x i1>
   %V4i64 = trunc <4 x i64> undef to <4 x i1>