[CostModel][X86] Improve single src shuffle costs

Add missing SK_PermuteSingleSrc costs for AVX2 targets and earlier, also added some of the simpler SK_PermuteTwoSrc costs to support splitting of SK_PermuteSingleSrc shuffles

llvm-svn: 310632

2 files changed, 96 insertions, 71 deletions

diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index 82425189a0a..fcacef84a66 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -838,6 +838,8 @@ int X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
     { TTI::SK_Alternate, MVT::v16i16, 1 }, // vpblendw
     { TTI::SK_Alternate, MVT::v32i8,  1 }, // vpblendvb
 
+    { TTI::SK_PermuteSingleSrc, MVT::v4f64,  1 }, // vpermpd
+    { TTI::SK_PermuteSingleSrc, MVT::v8f32,  1 }, // vpermps
     { TTI::SK_PermuteSingleSrc, MVT::v4i64,  1 }, // vpermq
     { TTI::SK_PermuteSingleSrc, MVT::v8i32,  1 }, // vpermd
     { TTI::SK_PermuteSingleSrc, MVT::v16i16, 4 }, // vperm2i128 + 2 * vpshufb
@@ -872,7 +874,16 @@ int X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
     { TTI::SK_Alternate, MVT::v8i32,  1 }, // vblendps
     { TTI::SK_Alternate, MVT::v8f32,  1 }, // vblendps
     { TTI::SK_Alternate, MVT::v16i16, 3 }, // vpand + vpandn + vpor
-    { TTI::SK_Alternate, MVT::v32i8,  3 }  // vpand + vpandn + vpor
+    { TTI::SK_Alternate, MVT::v32i8,  3 }, // vpand + vpandn + vpor
+
+    { TTI::SK_PermuteSingleSrc, MVT::v4f64,  3 }, // 2*vperm2f128 + vshufpd
+    { TTI::SK_PermuteSingleSrc, MVT::v4i64,  3 }, // 2*vperm2f128 + vshufpd
+    { TTI::SK_PermuteSingleSrc, MVT::v8f32,  4 }, // 2*vperm2f128 + 2*vshufps
+    { TTI::SK_PermuteSingleSrc, MVT::v8i32,  4 }, // 2*vperm2f128 + 2*vshufps
+    { TTI::SK_PermuteSingleSrc, MVT::v16i16, 8 }, // vextractf128 + 4*pshufb
+                                                  // + 2*por + vinsertf128
+    { TTI::SK_PermuteSingleSrc, MVT::v32i8,  8 }, // vextractf128 + 4*pshufb
+                                                  // + 2*por + vinsertf128
   };
 
   if (ST->hasAVX())
@@ -899,11 +910,14 @@ int X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
     { TTI::SK_Reverse,   MVT::v8i16,  1 }, // pshufb
     { TTI::SK_Reverse,   MVT::v16i8,  1 }, // pshufb
 
-    { TTI::SK_Alternate, MVT::v8i16,  3 }, // pshufb + pshufb + por
-    { TTI::SK_Alternate, MVT::v16i8,  3 }, // pshufb + pshufb + por
+    { TTI::SK_Alternate, MVT::v8i16,  3 }, // 2*pshufb + por
+    { TTI::SK_Alternate, MVT::v16i8,  3 }, // 2*pshufb + por
 
     { TTI::SK_PermuteSingleSrc, MVT::v8i16, 1 }, // pshufb
-    { TTI::SK_PermuteSingleSrc, MVT::v16i8, 1 }  // pshufb
+    { TTI::SK_PermuteSingleSrc, MVT::v16i8, 1 }, // pshufb
+
+    { TTI::SK_PermuteTwoSrc,    MVT::v8i16, 3 }, // 2*pshufb + por
+    { TTI::SK_PermuteTwoSrc,    MVT::v16i8, 3 }, // 2*pshufb + por
   };
 
   if (ST->hasSSSE3())
@@ -914,13 +928,13 @@ int X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
     { TTI::SK_Broadcast, MVT::v2f64,  1 }, // shufpd
     { TTI::SK_Broadcast, MVT::v2i64,  1 }, // pshufd
     { TTI::SK_Broadcast, MVT::v4i32,  1 }, // pshufd
-    { TTI::SK_Broadcast, MVT::v8i16,  2 }, // pshuflw  + pshufd
+    { TTI::SK_Broadcast, MVT::v8i16,  2 }, // pshuflw + pshufd
     { TTI::SK_Broadcast, MVT::v16i8,  3 }, // unpck + pshuflw + pshufd
 
     { TTI::SK_Reverse,   MVT::v2f64,  1 }, // shufpd
     { TTI::SK_Reverse,   MVT::v2i64,  1 }, // pshufd
     { TTI::SK_Reverse,   MVT::v4i32,  1 }, // pshufd
-    { TTI::SK_Reverse,   MVT::v8i16,  3 }, // pshuflw + pshufhw  + pshufd
+    { TTI::SK_Reverse,   MVT::v8i16,  3 }, // pshuflw + pshufhw + pshufd
     { TTI::SK_Reverse,   MVT::v16i8,  9 }, // 2*pshuflw + 2*pshufhw
                                            // + 2*pshufd + 2*unpck + packus
 
@@ -930,8 +944,17 @@ int X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
     { TTI::SK_Alternate, MVT::v8i16,  3 }, // pand + pandn + por
     { TTI::SK_Alternate, MVT::v16i8,  3 }, // pand + pandn + por
 
-    { TTI::SK_PermuteSingleSrc, MVT::v2i64, 1 }, // pshufd
-    { TTI::SK_PermuteSingleSrc, MVT::v4i32, 1 }  // pshufd
+    { TTI::SK_PermuteSingleSrc, MVT::v2f64,  1 }, // shufpd
+    { TTI::SK_PermuteSingleSrc, MVT::v2i64,  1 }, // pshufd
+    { TTI::SK_PermuteSingleSrc, MVT::v4i32,  1 }, // pshufd
+    { TTI::SK_PermuteSingleSrc, MVT::v8i16,  5 }, // 2*pshuflw + 2*pshufhw
+                                                  // + pshufd/unpck
+    { TTI::SK_PermuteSingleSrc, MVT::v16i8, 10 }, // 2*pshuflw + 2*pshufhw
+                                                  // + 2*pshufd + 2*unpck + 2*packus
+
+    { TTI::SK_PermuteTwoSrc,    MVT::v2f64,  1 }, // shufpd
+    { TTI::SK_PermuteTwoSrc,    MVT::v2i64,  1 }, // shufpd
+    { TTI::SK_PermuteTwoSrc,    MVT::v4i32,  2 }, // 2*{unpck,movsd,pshufd}
   };
 
   if (ST->hasSSE2())
@@ -939,9 +962,11 @@ int X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
       return LT.first * Entry->Cost;
 
   static const CostTblEntry SSE1ShuffleTbl[] = {
-    { TTI::SK_Broadcast, MVT::v4f32,  1 }, // shufps
-    { TTI::SK_Reverse,   MVT::v4f32,  1 }, // shufps
-    { TTI::SK_Alternate, MVT::v4f32,  2 }  // 2*shufps
+    { TTI::SK_Broadcast,        MVT::v4f32, 1 }, // shufps
+    { TTI::SK_Reverse,          MVT::v4f32, 1 }, // shufps
+    { TTI::SK_Alternate,        MVT::v4f32, 2 }, // 2*shufps
+    { TTI::SK_PermuteSingleSrc, MVT::v4f32, 1 }, // shufps
+    { TTI::SK_PermuteTwoSrc,    MVT::v4f32, 2 }, // 2*shufps
   };
 
   if (ST->hasSSE1())
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-single-src.ll b/llvm/test/Analysis/CostModel/X86/shuffle-single-src.ll
index a4a0673bc3c..d63632d085e 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-single-src.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-single-src.ll
@@ -13,33 +13,33 @@
 ; CHECK-LABEL: 'test_vXf64'
 define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double> %src512, <16 x double> %src1024) {
 
-  ; SSE2: cost of 2 {{.*}} %V128 = shufflevector
-  ; SSSE3: cost of 2 {{.*}} %V128 = shufflevector
-  ; SSE42: cost of 2 {{.*}} %V128 = shufflevector
-  ; AVX1: cost of 2 {{.*}} %V128 = shufflevector
-  ; AVX2: cost of 2 {{.*}} %V128 = shufflevector
+  ; SSE2: cost of 1 {{.*}} %V128 = shufflevector
+  ; SSSE3: cost of 1 {{.*}} %V128 = shufflevector
+  ; SSE42: cost of 1 {{.*}} %V128 = shufflevector
+  ; AVX1: cost of 1 {{.*}} %V128 = shufflevector
+  ; AVX2: cost of 1 {{.*}} %V128 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V128 = shufflevector
   %V128 = shufflevector <2 x double> %src128, <2 x double> undef, <2 x i32> <i32 1, i32 1>
 
-  ; SSE2: cost of 4 {{.*}} %V256 = shufflevector
-  ; SSSE3: cost of 4 {{.*}} %V256 = shufflevector
-  ; SSE42: cost of 4 {{.*}} %V256 = shufflevector
-  ; AVX1: cost of 6 {{.*}} %V256 = shufflevector
-  ; AVX2: cost of 6 {{.*}} %V256 = shufflevector
+  ; SSE2: cost of 2 {{.*}} %V256 = shufflevector
+  ; SSSE3: cost of 2 {{.*}} %V256 = shufflevector
+  ; SSE42: cost of 2 {{.*}} %V256 = shufflevector
+  ; AVX1: cost of 3 {{.*}} %V256 = shufflevector
+  ; AVX2: cost of 1 {{.*}} %V256 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V256 = shufflevector
   %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 
-  ; SSE2: cost of 24 {{.*}} %V512 = shufflevector
-  ; SSSE3: cost of 24 {{.*}} %V512 = shufflevector
-  ; SSE42: cost of 24 {{.*}} %V512 = shufflevector
+  ; SSE2: cost of 12 {{.*}} %V512 = shufflevector
+  ; SSSE3: cost of 12 {{.*}} %V512 = shufflevector
+  ; SSE42: cost of 12 {{.*}} %V512 = shufflevector
   ; AVX1: cost of 12 {{.*}} %V512 = shufflevector
   ; AVX2: cost of 12 {{.*}} %V512 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V512 = shufflevector
   %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 
-  ; SSE2: cost of 112 {{.*}} %V1024 = shufflevector
-  ; SSSE3: cost of 112 {{.*}} %V1024 = shufflevector
-  ; SSE42: cost of 112 {{.*}} %V1024 = shufflevector
+  ; SSE2: cost of 56 {{.*}} %V1024 = shufflevector
+  ; SSSE3: cost of 56 {{.*}} %V1024 = shufflevector
+  ; SSE42: cost of 56 {{.*}} %V1024 = shufflevector
   ; AVX1: cost of 72 {{.*}} %V1024 = shufflevector
   ; AVX2: cost of 72 {{.*}} %V1024 = shufflevector
   ; AVX512: cost of 2 {{.*}} %V1024 = shufflevector
@@ -59,17 +59,17 @@ define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512)
   ; AVX512: cost of 1 {{.*}} %V128 = shufflevector
   %V128 = shufflevector <2 x i64> %src128, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
 
-  ; SSE2: cost of 8 {{.*}} %V256 = shufflevector
-  ; SSSE3: cost of 8 {{.*}} %V256 = shufflevector
-  ; SSE42: cost of 8 {{.*}} %V256 = shufflevector
-  ; AVX1: cost of 8 {{.*}} %V256 = shufflevector
+  ; SSE2: cost of 2 {{.*}} %V256 = shufflevector
+  ; SSSE3: cost of 2 {{.*}} %V256 = shufflevector
+  ; SSE42: cost of 2 {{.*}} %V256 = shufflevector
+  ; AVX1: cost of 3 {{.*}} %V256 = shufflevector
   ; AVX2: cost of 1 {{.*}} %V256 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V256 = shufflevector
   %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 
-  ; SSE2: cost of 48 {{.*}} %V512 = shufflevector
-  ; SSSE3: cost of 48 {{.*}} %V512 = shufflevector
-  ; SSE42: cost of 48 {{.*}} %V512 = shufflevector
+  ; SSE2: cost of 12 {{.*}} %V512 = shufflevector
+  ; SSSE3: cost of 12 {{.*}} %V512 = shufflevector
+  ; SSE42: cost of 12 {{.*}} %V512 = shufflevector
   ; AVX1: cost of 16 {{.*}} %V512 = shufflevector
   ; AVX2: cost of 16 {{.*}} %V512 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V512 = shufflevector
@@ -81,25 +81,25 @@ define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512)
 ; CHECK-LABEL: 'test_vXf32'
 define void @test_vXf32(<4 x float> %src128, <8 x float> %src256, <16 x float> %src512) {
 
-  ; SSE2: cost of 6 {{.*}} %V128 = shufflevector
-  ; SSSE3: cost of 6 {{.*}} %V128 = shufflevector
-  ; SSE42: cost of 6 {{.*}} %V128 = shufflevector
-  ; AVX1: cost of 6 {{.*}} %V128 = shufflevector
-  ; AVX2: cost of 6 {{.*}} %V128 = shufflevector
+  ; SSE2: cost of 1 {{.*}} %V128 = shufflevector
+  ; SSSE3: cost of 1 {{.*}} %V128 = shufflevector
+  ; SSE42: cost of 1 {{.*}} %V128 = shufflevector
+  ; AVX1: cost of 1 {{.*}} %V128 = shufflevector
+  ; AVX2: cost of 1 {{.*}} %V128 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V128 = shufflevector
   %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 
-  ; SSE2: cost of 12 {{.*}} %V256 = shufflevector
-  ; SSSE3: cost of 12 {{.*}} %V256 = shufflevector
-  ; SSE42: cost of 12 {{.*}} %V256 = shufflevector
-  ; AVX1: cost of 14 {{.*}} %V256 = shufflevector
-  ; AVX2: cost of 14 {{.*}} %V256 = shufflevector
+  ; SSE2: cost of 4 {{.*}} %V256 = shufflevector
+  ; SSSE3: cost of 4 {{.*}} %V256 = shufflevector
+  ; SSE42: cost of 4 {{.*}} %V256 = shufflevector
+  ; AVX1: cost of 4 {{.*}} %V256 = shufflevector
+  ; AVX2: cost of 1 {{.*}} %V256 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V256 = shufflevector
   %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 
-  ; SSE2: cost of 72 {{.*}} %V512 = shufflevector
-  ; SSSE3: cost of 72 {{.*}} %V512 = shufflevector
-  ; SSE42: cost of 72 {{.*}} %V512 = shufflevector
+  ; SSE2: cost of 24 {{.*}} %V512 = shufflevector
+  ; SSSE3: cost of 24 {{.*}} %V512 = shufflevector
+  ; SSE42: cost of 24 {{.*}} %V512 = shufflevector
   ; AVX1: cost of 28 {{.*}} %V512 = shufflevector
   ; AVX2: cost of 28 {{.*}} %V512 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V512 = shufflevector
@@ -119,25 +119,25 @@ define void @test_vXi32(<4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512
   ; AVX512: cost of 1 {{.*}} %V128 = shufflevector
   %V128 = shufflevector <4 x i32> %src128, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 
-  ; SSE2: cost of 16 {{.*}} %V256 = shufflevector
-  ; SSSE3: cost of 16 {{.*}} %V256 = shufflevector
-  ; SSE42: cost of 16 {{.*}} %V256 = shufflevector
-  ; AVX1: cost of 16 {{.*}} %V256 = shufflevector
+  ; SSE2: cost of 4 {{.*}} %V256 = shufflevector
+  ; SSSE3: cost of 4 {{.*}} %V256 = shufflevector
+  ; SSE42: cost of 4 {{.*}} %V256 = shufflevector
+  ; AVX1: cost of 4 {{.*}} %V256 = shufflevector
   ; AVX2: cost of 1 {{.*}} %V256 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V256 = shufflevector
   %V256 = shufflevector <8 x i32> %src256, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 5, i32 3, i32 2, i32 1, i32 0>
 
-  ; SSE2: cost of 96 {{.*}} %V512 = shufflevector
-  ; SSSE3: cost of 96 {{.*}} %V512 = shufflevector
-  ; SSE42: cost of 96 {{.*}} %V512 = shufflevector
+  ; SSE2: cost of 24 {{.*}} %V512 = shufflevector
+  ; SSSE3: cost of 24 {{.*}} %V512 = shufflevector
+  ; SSE42: cost of 24 {{.*}} %V512 = shufflevector
   ; AVX1: cost of 32 {{.*}} %V512 = shufflevector
   ; AVX2: cost of 32 {{.*}} %V512 = shufflevector
   ; AVX512: cost of 1 {{.*}} %V512 = shufflevector
   %V512 = shufflevector <16 x i32> %src512, <16 x i32> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 
-  ; SSE2: cost of 448 {{.*}} %V1024 = shufflevector
-  ; SSSE3: cost of 448 {{.*}} %V1024 = shufflevector
-  ; SSE42: cost of 448 {{.*}} %V1024 = shufflevector
+  ; SSE2: cost of 112 {{.*}} %V1024 = shufflevector
+  ; SSSE3: cost of 112 {{.*}} %V1024 = shufflevector
+  ; SSE42: cost of 112 {{.*}} %V1024 = shufflevector
   ; AVX1: cost of 192 {{.*}} %V1024 = shufflevector
   ; AVX2: cost of 192 {{.*}} %V1024 = shufflevector
   ; AVX512: cost of 2 {{.*}} %V1024 = shufflevector
@@ -148,7 +148,7 @@ define void @test_vXi32(<4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512
 ; CHECK-LABEL: 'test_vXi16'
 define void @test_vXi16(<8 x i16> %src128, <16 x i16> %src256, <32 x i16> %src512, <64 x i16> %src1024) {
 
-  ; SSE2: cost of 16 {{.*}} %V128 = shufflevector
+  ; SSE2: cost of 5 {{.*}} %V128 = shufflevector
   ; SSSE3: cost of 1 {{.*}} %V128 = shufflevector
   ; SSE42: cost of 1 {{.*}} %V128 = shufflevector
   ; AVX1: cost of 1 {{.*}} %V128 = shufflevector
@@ -158,17 +158,17 @@ define void @test_vXi16(<8 x i16> %src128, <16 x i16> %src256, <32 x i16> %src51
   %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 
   ; SSE2: cost of 32 {{.*}} %V256 = shufflevector
-  ; SSSE3: cost of 32 {{.*}} %V256 = shufflevector
-  ; SSE42: cost of 32 {{.*}} %V256 = shufflevector
-  ; AVX1: cost of 32 {{.*}} %V256 = shufflevector
+  ; SSSE3: cost of 6 {{.*}} %V256 = shufflevector
+  ; SSE42: cost of 6 {{.*}} %V256 = shufflevector
+  ; AVX1: cost of 8 {{.*}} %V256 = shufflevector
   ; AVX2: cost of 4 {{.*}} %V256 = shufflevector
   ; AVX512F: cost of 4 {{.*}} %V256 = shufflevector
   ; AVX512BW cost of 1 {{.*}} %V256 = shufflevector
   %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 
   ; SSE2: cost of 192 {{.*}} %V512 = shufflevector
-  ; SSSE3: cost of 192 {{.*}} %V512 = shufflevector
-  ; SSE42: cost of 192 {{.*}} %V512 = shufflevector
+  ; SSSE3: cost of 36 {{.*}} %V512 = shufflevector
+  ; SSE42: cost of 36 {{.*}} %V512 = shufflevector
   ; AVX1: cost of 64 {{.*}} %V512 = shufflevector
   ; AVX2: cost of 64 {{.*}} %V512 = shufflevector
   ; AVX512F: cost of 64 {{.*}} %V512 = shufflevector
@@ -176,8 +176,8 @@ define void @test_vXi16(<8 x i16> %src128, <16 x i16> %src256, <32 x i16> %src51
   %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 
   ; SSE2: cost of 896 {{.*}} %V1024 = shufflevector
-  ; SSSE3: cost of 896 {{.*}} %V1024 = shufflevector
-  ; SSE42: cost of 896 {{.*}} %V1024 = shufflevector
+  ; SSSE3: cost of 168 {{.*}} %V1024 = shufflevector
+  ; SSE42: cost of 168 {{.*}} %V1024 = shufflevector
   ; AVX1: cost of 384 {{.*}} %V1024 = shufflevector
   ; AVX2: cost of 384 {{.*}} %V1024 = shufflevector
   ; AVX512F: cost of 384 {{.*}} %V1024 = shufflevector
@@ -188,7 +188,7 @@ define void @test_vXi16(<8 x i16> %src128, <16 x i16> %src256, <32 x i16> %src51
 
 ; CHECK-LABEL: 'test_vXi8'
 define void @test_vXi8(<16 x i8> %src128, <32 x i8> %src256, <64 x i8> %src512) {
-  ; SSE2: cost of 32 {{.*}} %V128 = shufflevector
+  ; SSE2: cost of 10 {{.*}} %V128 = shufflevector
   ; SSSE3: cost of 1 {{.*}} %V128 = shufflevector
   ; SSE42: cost of 1 {{.*}} %V128 = shufflevector
   ; AVX1: cost of 1 {{.*}} %V128 = shufflevector
@@ -197,17 +197,17 @@ define void @test_vXi8(<16 x i8> %src128, <32 x i8> %src256, <64 x i8> %src512)
   %V128 = shufflevector <16 x i8> %src128, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 
   ; SSE2: cost of 64 {{.*}} %V256 = shufflevector
-  ; SSSE3: cost of 64 {{.*}} %V256 = shufflevector
-  ; SSE42: cost of 64 {{.*}} %V256 = shufflevector
-  ; AVX1: cost of 64 {{.*}} %V256 = shufflevector
+  ; SSSE3: cost of 6 {{.*}} %V256 = shufflevector
+  ; SSE42: cost of 6 {{.*}} %V256 = shufflevector
+  ; AVX1: cost of 8 {{.*}} %V256 = shufflevector
   ; AVX2: cost of 4 {{.*}} %V256 = shufflevector
   ; AVX512F: cost of 4 {{.*}} %V256 = shufflevector
   ; AVX512BW: cost of 3 {{.*}} %V256 = shufflevector
   %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
 
   ; SSE2: cost of 384 {{.*}} %V512 = shufflevector
-  ; SSSE3: cost of 384 {{.*}} %V512 = shufflevector
-  ; SSE42: cost of 384 {{.*}} %V512 = shufflevector
+  ; SSSE3: cost of 36 {{.*}} %V512 = shufflevector
+  ; SSE42: cost of 36 {{.*}} %V512 = shufflevector
   ; AVX1: cost of 128 {{.*}} %V512 = shufflevector
   ; AVX2: cost of 128 {{.*}} %V512 = shufflevector
   ; AVX512F: cost of 128 {{.*}} %V512 = shufflevector