Skip to content

[CostModel][X86] Update SK_Reverse based on cost kinds #150650

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 2 commits into from
Jul 26, 2025
Merged

[CostModel][X86] Update SK_Reverse based on cost kinds #150650

merged 2 commits into from
Jul 26, 2025

Conversation

RKSimon
Copy link
Collaborator

@RKSimon RKSimon commented Jul 25, 2025

When these were converted to CostKindTblEntry the throughput was mainly copied to all cost kinds

Regenerated with my check_cost_tables.py helper script

@RKSimon
[CostModel][X86] Update SK_Reverse based on cost kinds
ab96552 
When these were converted to CostKindTblEntry the throughput was mainly copied to all cost kinds

Regenerated with my check_cost_tables.py helper script
@llvmbot llvmbot added backend:X86 llvm:analysis Includes value tracking, cost tables and constant folding labels Jul 25, 2025
@llvmbot
Copy link
Member

llvmbot commented Jul 25, 2025 •
edited
Loading

@llvm/pr-subscribers-backend-x86

@llvm/pr-subscribers-llvm-analysis

Author: Simon Pilgrim (RKSimon)

Changes

When these were converted to CostKindTblEntry the throughput was mainly copied to all cost kinds

Regenerated with my check_cost_tables.py helper script

Patch is 72.57 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/150650.diff

3 Files Affected:

  • (modified) llvm/lib/Target/X86/X86TargetTransformInfo.cpp (+28-27)
  • (modified) llvm/test/Analysis/CostModel/X86/shuffle-reverse-fp16.ll (+4-4)
  • (modified) llvm/test/Analysis/CostModel/X86/shuffle-reverse.ll (+108-108)
diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index 37a7b37e4133a..0b24088b52792 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -1842,10 +1842,11 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
     { TTI::SK_Broadcast, MVT::v32f16, { 1, 1, 1, 1 } }, // vpbroadcastw
     { TTI::SK_Broadcast, MVT::v64i8,  { 1, 1, 1, 1 } }, // vpbroadcastb
 
-    { TTI::SK_Reverse, MVT::v32i16,   { 2, 2, 2, 2 } }, // vpermw
-    { TTI::SK_Reverse, MVT::v32f16,   { 2, 2, 2, 2 } }, // vpermw
+    { TTI::SK_Reverse, MVT::v32i16,   { 2, 6, 2, 4 } }, // vpermw
+    { TTI::SK_Reverse, MVT::v32f16,   { 2, 6, 2, 4 } }, // vpermw
     { TTI::SK_Reverse, MVT::v16i16,   { 2, 2, 2, 2 } }, // vpermw
-    { TTI::SK_Reverse, MVT::v64i8,    { 2, 2, 2, 2 } }, // pshufb + vshufi64x2
+    { TTI::SK_Reverse, MVT::v16f16,   { 2, 2, 2, 2 } }, // vpermw
+    { TTI::SK_Reverse, MVT::v64i8,    { 2, 9, 2, 3 } }, // pshufb + vshufi64x2
 
     { TTI::SK_PermuteSingleSrc, MVT::v32i16, { 2, 2, 2, 2 } }, // vpermw
     { TTI::SK_PermuteSingleSrc, MVT::v32f16, { 2, 2, 2, 2 } }, // vpermw
@@ -1882,10 +1883,10 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
       {TTI::SK_Broadcast, MVT::v32f16, { 1, 1, 1, 1 } }, // vpbroadcastw
       {TTI::SK_Broadcast, MVT::v64i8,  { 1, 1, 1, 1 } }, // vpbroadcastb
 
-      {TTI::SK_Reverse, MVT::v8f64,  { 1, 3, 1, 1 } }, // vpermpd
-      {TTI::SK_Reverse, MVT::v16f32, { 1, 3, 1, 1 } }, // vpermps
-      {TTI::SK_Reverse, MVT::v8i64,  { 1, 3, 1, 1 } }, // vpermq
-      {TTI::SK_Reverse, MVT::v16i32, { 1, 3, 1, 1 } }, // vpermd
+      {TTI::SK_Reverse, MVT::v8f64,  { 1, 5, 2, 3 } }, // vpermpd
+      {TTI::SK_Reverse, MVT::v16f32, { 1, 3, 2, 3 } }, // vpermps
+      {TTI::SK_Reverse, MVT::v8i64,  { 1, 5, 2, 3 } }, // vpermq
+      {TTI::SK_Reverse, MVT::v16i32, { 1, 3, 2, 3 } }, // vpermd
       {TTI::SK_Reverse, MVT::v32i16, { 7, 7, 7, 7 } }, // per mca
       {TTI::SK_Reverse, MVT::v32f16, { 7, 7, 7, 7 } }, // per mca
       {TTI::SK_Reverse, MVT::v64i8,  { 7, 7, 7, 7 } }, // per mca
@@ -1981,13 +1982,13 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
     { TTI::SK_Broadcast, MVT::v16f16, { 1, 1, 1, 1 } }, // vpbroadcastw
     { TTI::SK_Broadcast, MVT::v32i8,  { 1, 1, 1, 1 } }, // vpbroadcastb
 
-    { TTI::SK_Reverse, MVT::v4f64,    { 1, 1, 1, 1 } }, // vpermpd
-    { TTI::SK_Reverse, MVT::v8f32,    { 1, 1, 1, 1 } }, // vpermps
-    { TTI::SK_Reverse, MVT::v4i64,    { 1, 1, 1, 1 } }, // vpermq
-    { TTI::SK_Reverse, MVT::v8i32,    { 1, 1, 1, 1 } }, // vpermd
-    { TTI::SK_Reverse, MVT::v16i16,   { 2, 2, 2, 2 } }, // vperm2i128 + pshufb
-    { TTI::SK_Reverse, MVT::v16f16,   { 2, 2, 2, 2 } }, // vperm2i128 + pshufb
-    { TTI::SK_Reverse, MVT::v32i8,    { 2, 2, 2, 2 } }, // vperm2i128 + pshufb
+    { TTI::SK_Reverse, MVT::v4f64,    { 1, 6, 1, 2 } }, // vpermpd
+    { TTI::SK_Reverse, MVT::v8f32,    { 2, 7, 2, 4 } }, // vpermps
+    { TTI::SK_Reverse, MVT::v4i64,    { 1, 6, 1, 2 } }, // vpermq
+    { TTI::SK_Reverse, MVT::v8i32,    { 2, 7, 2, 4 } }, // vpermd
+    { TTI::SK_Reverse, MVT::v16i16,   { 2, 9, 2, 4 } }, // vperm2i128 + pshufb
+    { TTI::SK_Reverse, MVT::v16f16,   { 2, 9, 2, 4 } }, // vperm2i128 + pshufb
+    { TTI::SK_Reverse, MVT::v32i8,    { 2, 9, 2, 4 } }, // vperm2i128 + pshufb
 
     { TTI::SK_Select, MVT::v16i16,    { 1, 1, 1, 1 } }, // vpblendvb
     { TTI::SK_Select, MVT::v16f16,    { 1, 1, 1, 1 } }, // vpblendvb
@@ -2085,15 +2086,15 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
       {TTI::SK_Broadcast, MVT::v16f16, {3,3,3,3}}, // vpshuflw + vpshufd + vinsertf128
       {TTI::SK_Broadcast, MVT::v32i8,  {2,2,2,2}}, // vpshufb + vinsertf128
 
-      {TTI::SK_Reverse, MVT::v4f64,  {2,2,2,2}}, // vperm2f128 + vpermilpd
-      {TTI::SK_Reverse, MVT::v8f32,  {2,2,2,2}}, // vperm2f128 + vpermilps
-      {TTI::SK_Reverse, MVT::v4i64,  {2,2,2,2}}, // vperm2f128 + vpermilpd
-      {TTI::SK_Reverse, MVT::v8i32,  {2,2,2,2}}, // vperm2f128 + vpermilps
-      {TTI::SK_Reverse, MVT::v16i16, {4,4,4,4}}, // vextractf128 + 2*pshufb
+      {TTI::SK_Reverse, MVT::v4f64,  {2,6,2,2}}, // vperm2f128 + vpermilpd
+      {TTI::SK_Reverse, MVT::v8f32,  {2,7,2,4}}, // vperm2f128 + vpermilps
+      {TTI::SK_Reverse, MVT::v4i64,  {2,6,2,2}}, // vperm2f128 + vpermilpd
+      {TTI::SK_Reverse, MVT::v8i32,  {2,7,2,4}}, // vperm2f128 + vpermilps
+      {TTI::SK_Reverse, MVT::v16i16, {2,9,5,5}}, // vextractf128 + 2*pshufb
                                                  // + vinsertf128
-      {TTI::SK_Reverse, MVT::v16f16, {4,4,4,4}}, // vextractf128 + 2*pshufb
+      {TTI::SK_Reverse, MVT::v16f16, {2,9,5,5}}, // vextractf128 + 2*pshufb
                                                  // + vinsertf128
-      {TTI::SK_Reverse, MVT::v32i8,  {4,4,4,4}}, // vextractf128 + 2*pshufb
+      {TTI::SK_Reverse, MVT::v32i8,  {2,9,5,5}}, // vextractf128 + 2*pshufb
                                                  // + vinsertf128
 
       {TTI::SK_Select, MVT::v4i64,  {1,1,1,1}}, // vblendpd
@@ -2160,9 +2161,9 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
       {TTI::SK_Broadcast, MVT::v8f16, {1, 1, 1, 1}}, // pshufb
       {TTI::SK_Broadcast, MVT::v16i8, {1, 1, 1, 1}}, // pshufb
 
-      {TTI::SK_Reverse, MVT::v8i16, {1, 1, 1, 1}}, // pshufb
-      {TTI::SK_Reverse, MVT::v8f16, {1, 1, 1, 1}}, // pshufb
-      {TTI::SK_Reverse, MVT::v16i8, {1, 1, 1, 1}}, // pshufb
+      {TTI::SK_Reverse, MVT::v8i16, {1, 2, 1, 2}}, // pshufb
+      {TTI::SK_Reverse, MVT::v8f16, {1, 2, 1, 2}}, // pshufb
+      {TTI::SK_Reverse, MVT::v16i8, {1, 2, 1, 2}}, // pshufb
 
       {TTI::SK_Select, MVT::v8i16, {3, 3, 3, 3}}, // 2*pshufb + por
       {TTI::SK_Select, MVT::v8f16, {3, 3, 3, 3}}, // 2*pshufb + por
@@ -2199,9 +2200,9 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
       {TTI::SK_Reverse, MVT::v2f64, {1, 1, 1, 1}}, // shufpd
       {TTI::SK_Reverse, MVT::v2i64, {1, 1, 1, 1}}, // pshufd
       {TTI::SK_Reverse, MVT::v4i32, {1, 1, 1, 1}}, // pshufd
-      {TTI::SK_Reverse, MVT::v8i16, {3, 3, 3, 3}}, // pshuflw + pshufhw + pshufd
-      {TTI::SK_Reverse, MVT::v8f16, {3, 3, 3, 3}}, // pshuflw + pshufhw + pshufd
-      {TTI::SK_Reverse, MVT::v16i8, {9, 9, 9, 9}}, // 2*pshuflw + 2*pshufhw
+      {TTI::SK_Reverse, MVT::v8i16, {2, 3, 3, 3}}, // pshuflw + pshufhw + pshufd
+      {TTI::SK_Reverse, MVT::v8f16, {2, 3, 3, 3}}, // pshuflw + pshufhw + pshufd
+      {TTI::SK_Reverse, MVT::v16i8, {5, 6,11,11}}, // 2*pshuflw + 2*pshufhw
                                                    // + 2*pshufd + 2*unpck + packus
 
       {TTI::SK_Select, MVT::v2i64, {1, 1, 1, 1}}, // movsd
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-reverse-fp16.ll b/llvm/test/Analysis/CostModel/X86/shuffle-reverse-fp16.ll
index 4affc7e97f6e5..f47e1a35aa087 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-reverse-fp16.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-reverse-fp16.ll
@@ -3,11 +3,11 @@
 
 define void @test_vXf16(<2 x half> %src32, <4 x half> %src64, <8 x half> %src128, <16 x half> %src256, <32 x half> %src512) {
 ; CHECK-LABEL: 'test_vXf16'
-; CHECK-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <2 x half> %src32, <2 x half> undef, <2 x i32> <i32 1, i32 0>
-; CHECK-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <4 x half> %src64, <4 x half> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <8 x half> %src128, <8 x half> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:2 SizeLat:2 for: %V32 = shufflevector <2 x half> %src32, <2 x half> undef, <2 x i32> <i32 1, i32 0>
+; CHECK-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:2 SizeLat:2 for: %V64 = shufflevector <4 x half> %src64, <4 x half> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:2 SizeLat:2 for: %V128 = shufflevector <8 x half> %src128, <8 x half> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found costs of 2 for: %V256 = shufflevector <16 x half> %src256, <16 x half> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:  Cost Model: Found costs of 2 for: %V512 = shufflevector <32 x half> %src512, <32 x half> 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 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:  Cost Model: Found costs of RThru:2 CodeSize:2 Lat:6 SizeLat:4 for: %V512 = shufflevector <32 x half> %src512, <32 x half> 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 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
   %V32  = shufflevector <2 x half> %src32, <2 x half> undef, <2 x i32> <i32 1, i32 0>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-reverse.ll b/llvm/test/Analysis/CostModel/X86/shuffle-reverse.ll
index 1bcf2d51f098b..aca0ba3ca74e4 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-reverse.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-reverse.ll
@@ -25,20 +25,20 @@ define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double>
 ;
 ; AVX1-LABEL: 'test_vXf64'
 ; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x double> %src128, <2 x double> undef, <2 x i32> <i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found costs of 2 for: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found costs of 4 for: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found costs of RThru:2 CodeSize:2 Lat:6 SizeLat:2 for: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found costs of RThru:4 CodeSize:4 Lat:12 SizeLat:4 for: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX2-LABEL: 'test_vXf64'
 ; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x double> %src128, <2 x double> undef, <2 x i32> <i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found costs of 2 for: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:6 SizeLat:2 for: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found costs of RThru:2 CodeSize:2 Lat:12 SizeLat:4 for: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX512-LABEL: 'test_vXf64'
 ; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x double> %src128, <2 x double> undef, <2 x i32> <i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:3 SizeLat:1 for: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:6 SizeLat:2 for: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:2 Lat:5 SizeLat:3 for: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
   %V128 = shufflevector <2 x double> %src128, <2 x double> undef, <2 x i32> <i32 1, i32 0>
@@ -56,20 +56,20 @@ define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512)
 ;
 ; AVX1-LABEL: 'test_vXi64'
 ; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x i64> %src128, <2 x i64> undef, <2 x i32> <i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found costs of 2 for: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found costs of 4 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found costs of RThru:2 CodeSize:2 Lat:6 SizeLat:2 for: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found costs of RThru:4 CodeSize:4 Lat:12 SizeLat:4 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX2-LABEL: 'test_vXi64'
 ; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x i64> %src128, <2 x i64> undef, <2 x i32> <i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found costs of 2 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:6 SizeLat:2 for: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found costs of RThru:2 CodeSize:2 Lat:12 SizeLat:4 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX512-LABEL: 'test_vXi64'
 ; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x i64> %src128, <2 x i64> undef, <2 x i32> <i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:3 SizeLat:1 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:6 SizeLat:2 for: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:2 Lat:5 SizeLat:3 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
   %V128 = shufflevector <2 x i64> %src128, <2 x i64> undef, <2 x i32> <i32 1, i32 0>
@@ -89,22 +89,22 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX1-LABEL: 'test_vXf32'
 ; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found costs of 2 for: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found costs of 4 for: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found costs of RThru:2 CodeSize:2 Lat:7 SizeLat:4 for: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found costs of RThru:4 CodeSize:4 Lat:14 SizeLat:8 for: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX2-LABEL: 'test_vXf32'
 ; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found costs of 2 for: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found costs of RThru:2 CodeSize:2 Lat:7 SizeLat:4 for: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found costs of RThru:4 CodeSize:4 Lat:14 SizeLat:8 for: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
 ; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:1 Lat:3 SizeLat:1 for: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found costs of RThru:2 CodeSize:2 Lat:7 SizeLat:4 for: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found costs of RThru:1 CodeSize:2 Lat:3 SizeLat:3 for: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
   %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 0>
@@ -125,22 +125,22 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX1-LABEL: 'test_vXi32'
 ; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector ...
[truncated]

@github-actions GitHub Actions
Copy link

github-actions bot commented Jul 25, 2025 •
edited
Loading

⚠️ C/C++ code formatter, clang-format found issues in your code. ⚠️

You can test this locally with the following command: 
git-clang-format --diff HEAD~1 HEAD --extensions cpp -- llvm/lib/Target/X86/X86TargetTransformInfo.cpp
View the diff from clang-format here. 
diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index 90791fc57..cf9557d19 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -1838,34 +1838,34 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
         return LT.first * *KindCost;
 
   static const CostKindTblEntry AVX512BWShuffleTbl[] = {
-    { TTI::SK_Broadcast, MVT::v32i16, { 1, 3, 1, 1 } }, // vpbroadcastw
-    { TTI::SK_Broadcast, MVT::v32f16, { 1, 3, 1, 1 } }, // vpbroadcastw
-    { TTI::SK_Broadcast, MVT::v64i8,  { 1, 3, 1, 1 } }, // vpbroadcastb
-
-    { TTI::SK_Reverse, MVT::v32i16,   { 2, 6, 2, 4 } }, // vpermw
-    { TTI::SK_Reverse, MVT::v32f16,   { 2, 6, 2, 4 } }, // vpermw
-    { TTI::SK_Reverse, MVT::v16i16,   { 2, 2, 2, 2 } }, // vpermw
-    { TTI::SK_Reverse, MVT::v16f16,   { 2, 2, 2, 2 } }, // vpermw
-    { TTI::SK_Reverse, MVT::v64i8,    { 2, 9, 2, 3 } }, // pshufb + vshufi64x2
-
-    { TTI::SK_PermuteSingleSrc, MVT::v32i16, { 2, 2, 2, 2 } }, // vpermw
-    { TTI::SK_PermuteSingleSrc, MVT::v32f16, { 2, 2, 2, 2 } }, // vpermw
-    { TTI::SK_PermuteSingleSrc, MVT::v16i16, { 2, 2, 2, 2 } }, // vpermw
-    { TTI::SK_PermuteSingleSrc, MVT::v16f16, { 2, 2, 2, 2 } }, // vpermw
-    { TTI::SK_PermuteSingleSrc, MVT::v64i8,  { 8, 8, 8, 8 } },  // extend to v32i16
-
-    { TTI::SK_PermuteTwoSrc, MVT::v32i16,{  2,  2,  2,  2 } }, // vpermt2w
-    { TTI::SK_PermuteTwoSrc, MVT::v32f16,{  2,  2,  2,  2 } }, // vpermt2w
-    { TTI::SK_PermuteTwoSrc, MVT::v16i16,{  2,  2,  2,  2 } }, // vpermt2w
-    { TTI::SK_PermuteTwoSrc, MVT::v8i16, {  2,  2,  2,  2 } },  // vpermt2w
-    { TTI::SK_PermuteTwoSrc, MVT::v64i8, { 19, 19, 19, 19 } }, // 6 * v32i8 + 1
-
-    { TTI::SK_Select, MVT::v32i16, { 1, 1, 1, 1 } }, // vblendmw
-    { TTI::SK_Select, MVT::v64i8,  { 1, 1, 1, 1 } }, // vblendmb
-
-    { TTI::SK_Splice, MVT::v32i16, { 2, 2, 2, 2 } }, // vshufi64x2 + palignr
-    { TTI::SK_Splice, MVT::v32f16, { 2, 2, 2, 2 } }, // vshufi64x2 + palignr
-    { TTI::SK_Splice, MVT::v64i8,  { 2, 2, 2, 2 } }, // vshufi64x2 + palignr
+      {TTI::SK_Broadcast, MVT::v32i16, {1, 3, 1, 1}}, // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v32f16, {1, 3, 1, 1}}, // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v64i8, {1, 3, 1, 1}},  // vpbroadcastb
+
+      {TTI::SK_Reverse, MVT::v32i16, {2, 6, 2, 4}}, // vpermw
+      {TTI::SK_Reverse, MVT::v32f16, {2, 6, 2, 4}}, // vpermw
+      {TTI::SK_Reverse, MVT::v16i16, {2, 2, 2, 2}}, // vpermw
+      {TTI::SK_Reverse, MVT::v16f16, {2, 2, 2, 2}}, // vpermw
+      {TTI::SK_Reverse, MVT::v64i8, {2, 9, 2, 3}},  // pshufb + vshufi64x2
+
+      {TTI::SK_PermuteSingleSrc, MVT::v32i16, {2, 2, 2, 2}}, // vpermw
+      {TTI::SK_PermuteSingleSrc, MVT::v32f16, {2, 2, 2, 2}}, // vpermw
+      {TTI::SK_PermuteSingleSrc, MVT::v16i16, {2, 2, 2, 2}}, // vpermw
+      {TTI::SK_PermuteSingleSrc, MVT::v16f16, {2, 2, 2, 2}}, // vpermw
+      {TTI::SK_PermuteSingleSrc, MVT::v64i8, {8, 8, 8, 8}},  // extend to v32i16
+
+      {TTI::SK_PermuteTwoSrc, MVT::v32i16, {2, 2, 2, 2}},    // vpermt2w
+      {TTI::SK_PermuteTwoSrc, MVT::v32f16, {2, 2, 2, 2}},    // vpermt2w
+      {TTI::SK_PermuteTwoSrc, MVT::v16i16, {2, 2, 2, 2}},    // vpermt2w
+      {TTI::SK_PermuteTwoSrc, MVT::v8i16, {2, 2, 2, 2}},     // vpermt2w
+      {TTI::SK_PermuteTwoSrc, MVT::v64i8, {19, 19, 19, 19}}, // 6 * v32i8 + 1
+
+      {TTI::SK_Select, MVT::v32i16, {1, 1, 1, 1}}, // vblendmw
+      {TTI::SK_Select, MVT::v64i8, {1, 1, 1, 1}},  // vblendmb
+
+      {TTI::SK_Splice, MVT::v32i16, {2, 2, 2, 2}}, // vshufi64x2 + palignr
+      {TTI::SK_Splice, MVT::v32f16, {2, 2, 2, 2}}, // vshufi64x2 + palignr
+      {TTI::SK_Splice, MVT::v64i8, {2, 2, 2, 2}},  // vshufi64x2 + palignr
   };
 
   if (ST->hasBWI())
@@ -1875,84 +1875,84 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
         return LT.first * *KindCost;
 
   static const CostKindTblEntry AVX512ShuffleTbl[] = {
-      {TTI::SK_Broadcast, MVT::v8f64,  { 1, 3, 1, 1 } }, // vbroadcastsd
-      {TTI::SK_Broadcast, MVT::v4f64,  { 1, 3, 1, 1 } }, // vbroadcastsd
-      {TTI::SK_Broadcast, MVT::v16f32, { 1, 3, 1, 1 } }, // vbroadcastss
-      {TTI::SK_Broadcast, MVT::v8f32,  { 1, 3, 1, 1 } }, // vbroadcastss
-      {TTI::SK_Broadcast, MVT::v8i64,  { 1, 3, 1, 1 } }, // vpbroadcastq
-      {TTI::SK_Broadcast, MVT::v4i64,  { 1, 3, 1, 1 } }, // vpbroadcastq
-      {TTI::SK_Broadcast, MVT::v16i32, { 1, 3, 1, 1 } }, // vpbroadcastd
-      {TTI::SK_Broadcast, MVT::v8i32,  { 1, 3, 1, 1 } }, // vpbroadcastd
-      {TTI::SK_Broadcast, MVT::v32i16, { 1, 3, 1, 1 } }, // vpbroadcastw
-      {TTI::SK_Broadcast, MVT::v16i16, { 1, 3, 1, 1 } }, // vpbroadcastw
-      {TTI::SK_Broadcast, MVT::v32f16, { 1, 3, 1, 1 } }, // vpbroadcastw
-      {TTI::SK_Broadcast, MVT::v16f16, { 1, 3, 1, 1 } }, // vpbroadcastw
-      {TTI::SK_Broadcast, MVT::v64i8,  { 1, 3, 1, 1 } }, // vpbroadcastb
-      {TTI::SK_Broadcast, MVT::v32i8,  { 1, 3, 1, 1 }},  // vpbroadcastb
-
-      {TTI::SK_Reverse, MVT::v8f64,  { 1, 5, 2, 3 } }, // vpermpd
-      {TTI::SK_Reverse, MVT::v16f32, { 1, 3, 2, 3 } }, // vpermps
-      {TTI::SK_Reverse, MVT::v8i64,  { 1, 5, 2, 3 } }, // vpermq
-      {TTI::SK_Reverse, MVT::v16i32, { 1, 3, 2, 3 } }, // vpermd
-      {TTI::SK_Reverse, MVT::v32i16, { 7, 7, 7, 7 } }, // per mca
-      {TTI::SK_Reverse, MVT::v32f16, { 7, 7, 7, 7 } }, // per mca
-      {TTI::SK_Reverse, MVT::v64i8,  { 7, 7, 7, 7 } }, // per mca
-
-      {TTI::SK_Splice, MVT::v8f64,  { 1, 1, 1, 1 } }, // vpalignd
-      {TTI::SK_Splice, MVT::v4f64,  { 1, 1, 1, 1 } }, // vpalignd
-      {TTI::SK_Splice, MVT::v16f32, { 1, 1, 1, 1 } }, // vpalignd
-      {TTI::SK_Splice, MVT::v8f32,  { 1, 1, 1, 1 } }, // vpalignd
-      {TTI::SK_Splice, MVT::v8i64,  { 1, 1, 1, 1 } }, // vpalignd
-      {TTI::SK_Splice, MVT::v4i64,  { 1, 1, 1, 1 } }, // vpalignd
-      {TTI::SK_Splice, MVT::v16i32, { 1, 1, 1, 1 } }, // vpalignd
-      {TTI::SK_Splice, MVT::v8i32,  { 1, 1, 1, 1 } }, // vpalignd
-      {TTI::SK_Splice, MVT::v32i16, { 4, 4, 4, 4 } }, // split + palignr
-      {TTI::SK_Splice, MVT::v32f16, { 4, 4, 4, 4 } }, // split + palignr
-      {TTI::SK_Splice, MVT::v64i8,  { 4, 4, 4, 4 } }, // split + palignr
-
-      {TTI::SK_PermuteSingleSrc, MVT::v8f64,  { 1, 3, 1, 1 } }, // vpermpd
-      {TTI::SK_PermuteSingleSrc, MVT::v4f64,  { 1, 3, 1, 1 } }, // vpermpd
-      {TTI::SK_PermuteSingleSrc, MVT::v2f64,  { 1, 3, 1, 1 } }, // vpermpd
-      {TTI::SK_PermuteSingleSrc, MVT::v16f32, { 1, 3, 1, 1 } }, // vpermps
-      {TTI::SK_PermuteSingleSrc, MVT::v8f32,  { 1, 3, 1, 1 } }, // vpermps
-      {TTI::SK_PermuteSingleSrc, MVT::v4f32,  { 1, 3, 1, 1 } }, // vpermps
-      {TTI::SK_PermuteSingleSrc, MVT::v8i64,  { 1, 3, 1, 1 } }, // vpermq
-      {TTI::SK_PermuteSingleSrc, MVT::v4i64,  { 1, 3, 1, 1 } }, // vpermq
-      {TTI::SK_PermuteSingleSrc, MVT::v2i64,  { 1, 3, 1, 1 } }, // vpermq
-      {TTI::SK_PermuteSingleSrc, MVT::v16i32, { 1, 3, 1, 1 } }, // vpermd
-      {TTI::SK_PermuteSingleSrc, MVT::v8i32,  { 1, 3, 1, 1 } }, // vpermd
-      {TTI::SK_PermuteSingleSrc, MVT::v4i32,  { 1, 3, 1, 1 } }, // vpermd
-      {TTI::SK_PermuteSingleSrc, MVT::v16i8,  { 1, 3, 1, 1 } }, // pshufb
-
-      {TTI::SK_PermuteTwoSrc, MVT::v8f64,  { 1, 3, 1, 1 } }, // vpermt2pd
-      {TTI::SK_PermuteTwoSrc, MVT::v16f32, { 1, 3, 1, 1 } }, // vpermt2ps
-      {TTI::SK_PermuteTwoSrc, MVT::v8i64,  { 1, 3, 1, 1 } }, // vpermt2q
-      {TTI::SK_PermuteTwoSrc, MVT::v16i32, { 1, 3, 1, 1 } }, // vpermt2d
-      {TTI::SK_PermuteTwoSrc, MVT::v4f64,  { 1, 3, 1, 1 } }, // vpermt2pd
-      {TTI::SK_PermuteTwoSrc, MVT::v8f32,  { 1, 3, 1, 1 } }, // vpermt2ps
-      {TTI::SK_PermuteTwoSrc, MVT::v4i64,  { 1, 3, 1, 1 } }, // vpermt2q
-      {TTI::SK_PermuteTwoSrc, MVT::v8i32,  { 1, 3, 1, 1 } }, // vpermt2d
-      {TTI::SK_PermuteTwoSrc, MVT::v2f64,  { 1, 3, 1, 1 } }, // vpermt2pd
-      {TTI::SK_PermuteTwoSrc, MVT::v4f32,  { 1, 3, 1, 1 } }, // vpermt2ps
-      {TTI::SK_PermuteTwoSrc, MVT::v2i64,  { 1, 3, 1, 1 } }, // vpermt2q
-      {TTI::SK_PermuteTwoSrc, MVT::v4i32,  { 1, 3, 1, 1 } }, // vpermt2d
+      {TTI::SK_Broadcast, MVT::v8f64, {1, 3, 1, 1}},  // vbroadcastsd
+      {TTI::SK_Broadcast, MVT::v4f64, {1, 3, 1, 1}},  // vbroadcastsd
+      {TTI::SK_Broadcast, MVT::v16f32, {1, 3, 1, 1}}, // vbroadcastss
+      {TTI::SK_Broadcast, MVT::v8f32, {1, 3, 1, 1}},  // vbroadcastss
+      {TTI::SK_Broadcast, MVT::v8i64, {1, 3, 1, 1}},  // vpbroadcastq
+      {TTI::SK_Broadcast, MVT::v4i64, {1, 3, 1, 1}},  // vpbroadcastq
+      {TTI::SK_Broadcast, MVT::v16i32, {1, 3, 1, 1}}, // vpbroadcastd
+      {TTI::SK_Broadcast, MVT::v8i32, {1, 3, 1, 1}},  // vpbroadcastd
+      {TTI::SK_Broadcast, MVT::v32i16, {1, 3, 1, 1}}, // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v16i16, {1, 3, 1, 1}}, // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v32f16, {1, 3, 1, 1}}, // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v16f16, {1, 3, 1, 1}}, // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v64i8, {1, 3, 1, 1}},  // vpbroadcastb
+      {TTI::SK_Broadcast, MVT::v32i8, {1, 3, 1, 1}},  // vpbroadcastb
+
+      {TTI::SK_Reverse, MVT::v8f64, {1, 5, 2, 3}},  // vpermpd
+      {TTI::SK_Reverse, MVT::v16f32, {1, 3, 2, 3}}, // vpermps
+      {TTI::SK_Reverse, MVT::v8i64, {1, 5, 2, 3}},  // vpermq
+      {TTI::SK_Reverse, MVT::v16i32, {1, 3, 2, 3}}, // vpermd
+      {TTI::SK_Reverse, MVT::v32i16, {7, 7, 7, 7}}, // per mca
+      {TTI::SK_Reverse, MVT::v32f16, {7, 7, 7, 7}}, // per mca
+      {TTI::SK_Reverse, MVT::v64i8, {7, 7, 7, 7}},  // per mca
+
+      {TTI::SK_Splice, MVT::v8f64, {1, 1, 1, 1}},  // vpalignd
+      {TTI::SK_Splice, MVT::v4f64, {1, 1, 1, 1}},  // vpalignd
+      {TTI::SK_Splice, MVT::v16f32, {1, 1, 1, 1}}, // vpalignd
+      {TTI::SK_Splice, MVT::v8f32, {1, 1, 1, 1}},  // vpalignd
+      {TTI::SK_Splice, MVT::v8i64, {1, 1, 1, 1}},  // vpalignd
+      {TTI::SK_Splice, MVT::v4i64, {1, 1, 1, 1}},  // vpalignd
+      {TTI::SK_Splice, MVT::v16i32, {1, 1, 1, 1}}, // vpalignd
+      {TTI::SK_Splice, MVT::v8i32, {1, 1, 1, 1}},  // vpalignd
+      {TTI::SK_Splice, MVT::v32i16, {4, 4, 4, 4}}, // split + palignr
+      {TTI::SK_Splice, MVT::v32f16, {4, 4, 4, 4}}, // split + palignr
+      {TTI::SK_Splice, MVT::v64i8, {4, 4, 4, 4}},  // split + palignr
+
+      {TTI::SK_PermuteSingleSrc, MVT::v8f64, {1, 3, 1, 1}},  // vpermpd
+      {TTI::SK_PermuteSingleSrc, MVT::v4f64, {1, 3, 1, 1}},  // vpermpd
+      {TTI::SK_PermuteSingleSrc, MVT::v2f64, {1, 3, 1, 1}},  // vpermpd
+      {TTI::SK_PermuteSingleSrc, MVT::v16f32, {1, 3, 1, 1}}, // vpermps
+      {TTI::SK_PermuteSingleSrc, MVT::v8f32, {1, 3, 1, 1}},  // vpermps
+      {TTI::SK_PermuteSingleSrc, MVT::v4f32, {1, 3, 1, 1}},  // vpermps
+      {TTI::SK_PermuteSingleSrc, MVT::v8i64, {1, 3, 1, 1}},  // vpermq
+      {TTI::SK_PermuteSingleSrc, MVT::v4i64, {1, 3, 1, 1}},  // vpermq
+      {TTI::SK_PermuteSingleSrc, MVT::v2i64, {1, 3, 1, 1}},  // vpermq
+      {TTI::SK_PermuteSingleSrc, MVT::v16i32, {1, 3, 1, 1}}, // vpermd
+      {TTI::SK_PermuteSingleSrc, MVT::v8i32, {1, 3, 1, 1}},  // vpermd
+      {TTI::SK_PermuteSingleSrc, MVT::v4i32, {1, 3, 1, 1}},  // vpermd
+      {TTI::SK_PermuteSingleSrc, MVT::v16i8, {1, 3, 1, 1}},  // pshufb
+
+      {TTI::SK_PermuteTwoSrc, MVT::v8f64, {1, 3, 1, 1}},  // vpermt2pd
+      {TTI::SK_PermuteTwoSrc, MVT::v16f32, {1, 3, 1, 1}}, // vpermt2ps
+      {TTI::SK_PermuteTwoSrc, MVT::v8i64, {1, 3, 1, 1}},  // vpermt2q
+      {TTI::SK_PermuteTwoSrc, MVT::v16i32, {1, 3, 1, 1}}, // vpermt2d
+      {TTI::SK_PermuteTwoSrc, MVT::v4f64, {1, 3, 1, 1}},  // vpermt2pd
+      {TTI::SK_PermuteTwoSrc, MVT::v8f32, {1, 3, 1, 1}},  // vpermt2ps
+      {TTI::SK_PermuteTwoSrc, MVT::v4i64, {1, 3, 1, 1}},  // vpermt2q
+      {TTI::SK_PermuteTwoSrc, MVT::v8i32, {1, 3, 1, 1}},  // vpermt2d
+      {TTI::SK_PermuteTwoSrc, MVT::v2f64, {1, 3, 1, 1}},  // vpermt2pd
+      {TTI::SK_PermuteTwoSrc, MVT::v4f32, {1, 3, 1, 1}},  // vpermt2ps
+      {TTI::SK_PermuteTwoSrc, MVT::v2i64, {1, 3, 1, 1}},  // vpermt2q
+      {TTI::SK_PermuteTwoSrc, MVT::v4i32, {1, 3, 1, 1}},  // vpermt2d
 
       // FIXME: This just applies the type legalization cost rules above
       // assuming these completely split.
-      {TTI::SK_PermuteSingleSrc, MVT::v32i16, { 14, 14, 14, 14 } },
-      {TTI::SK_PermuteSingleSrc, MVT::v32f16, { 14, 14, 14, 14 } },
-      {TTI::SK_PermuteSingleSrc, MVT::v64i8,  { 14, 14, 14, 14 } },
-      {TTI::SK_PermuteTwoSrc,    MVT::v32i16, { 42, 42, 42, 42 } },
-      {TTI::SK_PermuteTwoSrc,    MVT::v32f16, { 42, 42, 42, 42 } },
-      {TTI::SK_PermuteTwoSrc,    MVT::v64i8,  { 42, 42, 42, 42 } },
-
-      {TTI::SK_Select, MVT::v32i16, { 1, 1, 1, 1 } }, // vpternlogq
-      {TTI::SK_Select, MVT::v32f16, { 1, 1, 1, 1 } }, // vpternlogq
-      {TTI::SK_Select, MVT::v64i8,  { 1, 1, 1, 1 } }, // vpternlogq
-      {TTI::SK_Select, MVT::v8f64,  { 1, 1, 1, 1 } }, // vblendmpd
-      {TTI::SK_Select, MVT::v16f32, { 1, 1, 1, 1 } }, // vblendmps
-      {TTI::SK_Select, MVT::v8i64,  { 1, 1, 1, 1 } }, // vblendmq
-      {TTI::SK_Select, MVT::v16i32, { 1, 1, 1, 1 } }, // vblendmd
+      {TTI::SK_PermuteSingleSrc, MVT::v32i16, {14, 14, 14, 14}},
+      {TTI::SK_PermuteSingleSrc, MVT::v32f16, {14, 14, 14, 14}},
+      {TTI::SK_PermuteSingleSrc, MVT::v64i8, {14, 14, 14, 14}},
+      {TTI::SK_PermuteTwoSrc, MVT::v32i16, {42, 42, 42, 42}},
+      {TTI::SK_PermuteTwoSrc, MVT::v32f16, {42, 42, 42, 42}},
+      {TTI::SK_PermuteTwoSrc, MVT::v64i8, {42, 42, 42, 42}},
+
+      {TTI::SK_Select, MVT::v32i16, {1, 1, 1, 1}}, // vpternlogq
+      {TTI::SK_Select, MVT::v32f16, {1, 1, 1, 1}}, // vpternlogq
+      {TTI::SK_Select, MVT::v64i8, {1, 1, 1, 1}},  // vpternlogq
+      {TTI::SK_Select, MVT::v8f64, {1, 1, 1, 1}},  // vblendmpd
+      {TTI::SK_Select, MVT::v16f32, {1, 1, 1, 1}}, // vblendmps
+      {TTI::SK_Select, MVT::v8i64, {1, 1, 1, 1}},  // vblendmq
+      {TTI::SK_Select, MVT::v16i32, {1, 1, 1, 1}}, // vblendmd
   };
 
   if (ST->hasAVX512())
@@ -1981,50 +1981,50 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
         return LT.first * *KindCost;
 
   static const CostKindTblEntry AVX2ShuffleTbl[] = {
-    { TTI::SK_Broadcast, MVT::v4f64,  { 1, 3, 1, 2 } }, // vbroadcastpd
-    { TTI::SK_Broadcast, MVT::v8f32,  { 1, 3, 1, 2 } }, // vbroadcastps
-    { TTI::SK_Broadcast, MVT::v4i64,  { 1, 3, 1, 2 } }, // vpbroadcastq
-    { TTI::SK_Broadcast, MVT::v8i32,  { 1, 3, 1, 2 } }, // vpbroadcastd
-    { TTI::SK_Broadcast, MVT::v16i16, { 1, 3, 1, 2 } }, // vpbroadcastw
-    { TTI::SK_Broadcast, MVT::v8i16,  { 1, 3, 1, 1 } }, // vpbroadcastw
-    { TTI::SK_Broadcast, MVT::v16f16, { 1, 3, 1, 2 } }, // vpbroadcastw
-    { TTI::SK_Broadcast, MVT::v8f16,  { 1, 3, 1, 1 } }, // vpbroadcastw
-    { TTI::SK_Broadcast, MVT::v32i8,  { 1, 3, 1, 2 } }, // vpbroadcastb
-    { TTI::SK_Broadcast, MVT::v16i8,  { 1, 3, 1, 1 } }, // vpbroadcastb
-
-    { TTI::SK_Reverse, MVT::v4f64,    { 1, 6, 1, 2 } }, // vpermpd
-    { TTI::SK_Reverse, MVT::v8f32,    { 2, 7, 2, 4 } }, // vpermps
-    { TTI::SK_Reverse, MVT::v4i64,    { 1, 6, 1, 2 } }, // vpermq
-    { TTI::SK_Reverse, MVT::v8i32,    { 2, 7, 2, 4 } }, // vpermd
-    { TTI::SK_Reverse, MVT::v16i16,   { 2, 9, 2, 4 } }, // vperm2i128 + pshufb
-    { TTI::SK_Reverse, MVT::v16f16,   { 2, 9, 2, 4 } }, // vperm2i128 + pshufb
-    { TTI::SK_Reverse, MVT::v32i8,    { 2, 9, 2, 4 } }, // vperm2i128 + pshufb
-
-    { TTI::SK_Select, MVT::v16i16,    { 1, 1, 1, 1 } }, // vpblendvb
-    { TTI::SK_Select, MVT::v16f16,    { 1, 1, 1, 1 } }, // vpblendvb
-    { TTI::SK_Select, MVT::v32i8,     { 1, 1, 1, 1 } }, // vpblendvb
-
-    { TTI::SK_Splice, MVT::v8i32,     { 2, 2, 2, 2 } }, // vperm2i128 + vpalignr
-    { TTI::SK_Splice, MVT::v8f32,     { 2, 2, 2, 2 } }, // vperm2i128 + vpalignr
-    { TTI::SK_Splice, MVT::v16i16,    { 2, 2, 2, 2 } }, // vperm2i128 + vpalignr
-    { TTI::SK_Splice, MVT::v16f16,    { 2, 2, 2, 2 } }, // vperm2i128 + vpalignr
-    { TTI::SK_Splice, MVT::v32i8,     { 2, 2, 2, 2 } }, // vperm2i128 + vpalignr
-
-    { TTI::SK_PermuteSingleSrc, MVT::v4f64,  { 1, 1, 1, 1 } }, // vpermpd
-    { TTI::SK_PermuteSingleSrc, MVT::v8f32,  { 1, 1, 1, 1 } }, // vpermps
-    { TTI::SK_PermuteSingleSrc, MVT::v4i64,  { 1, 1, 1, 1 } }, // vpermq
-    { TTI::SK_PermuteSingleSrc, MVT::v8i32,  { 1, 1, 1, 1 } }, // vpermd
-    { TTI::SK_PermuteSingleSrc, MVT::v16i16, { 4, 4, 4, 4 } },
-    { TTI::SK_PermuteSingleSrc, MVT::v16f16, { 4, 4, 4, 4 } },
-    { TTI::SK_PermuteSingleSrc, MVT::v32i8,  { 4, 4, 4, 4 } },
-
-    { TTI::SK_PermuteTwoSrc, MVT::v4f64,  { 3, 3, 3, 3 } }, // 2*vpermpd + vblendpd
-    { TTI::SK_PermuteTwoSrc, MVT::v8f32,  { 3, 3, 3, 3 } }, // 2*vpermps + vblendps
-    { TTI::SK_PermuteTwoSrc, MVT::v4i64,  { 3, 3, 3, 3 } }, // 2*vpermq + vpblendd
-    { TTI::SK_PermuteTwoSrc, MVT::v8i32,  { 3, 3, 3, 3 } }, // 2*vpermd + vpblendd
-    { TTI::SK_PermuteTwoSrc, MVT::v16i16, { 7, 7, 7, 7 } },
-    { TTI::SK_PermuteTwoSrc, MVT::v16f16, { 7, 7, 7, 7 } },
-    { TTI::SK_PermuteTwoSrc, MVT::v32i8,  { 7, 7, 7, 7 } },
+      {TTI::SK_Broadcast, MVT::v4f64, {1, 3, 1, 2}},  // vbroadcastpd
+      {TTI::SK_Broadcast, MVT::v8f32, {1, 3, 1, 2}},  // vbroadcastps
+      {TTI::SK_Broadcast, MVT::v4i64, {1, 3, 1, 2}},  // vpbroadcastq
+      {TTI::SK_Broadcast, MVT::v8i32, {1, 3, 1, 2}},  // vpbroadcastd
+      {TTI::SK_Broadcast, MVT::v16i16, {1, 3, 1, 2}}, // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v8i16, {1, 3, 1, 1}},  // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v16f16, {1, 3, 1, 2}}, // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v8f16, {1, 3, 1, 1}},  // vpbroadcastw
+      {TTI::SK_Broadcast, MVT::v32i8, {1, 3, 1, 2}},  // vpbroadcastb
+      {TTI::SK_Broadcast, MVT::v16i8, {1, 3, 1, 1}},  // vpbroadcastb
+
+      {TTI::SK_Reverse, MVT::v4f64, {1, 6, 1, 2}},  // vpermpd
+      {TTI::SK_Reverse, MVT::v8f32, {2, 7, 2, 4}},  // vpermps
+      {TTI::SK_Reverse, MVT::v4i64, {1, 6, 1, 2}},  // vpermq
+      {TTI::SK_Reverse, MVT::v8i32, {2, 7, 2, 4}},  // vpermd
+      {TTI::SK_Reverse, MVT::v16i16, {2, 9, 2, 4}}, // vperm2i128 + pshufb
+      {TTI::SK_Reverse, MVT::v16f16, {2, 9, 2, 4}}, // vperm2i128 + pshufb
+      {TTI::SK_Reverse, MVT::v32i8, {2, 9, 2, 4}},  // vperm2i128 + pshufb
+
+      {TTI::SK_Select, MVT::v16i16, {1, 1, 1, 1}}, // vpblendvb
+      {TTI::SK_Select, MVT::v16f16, {1, 1, 1, 1}}, // vpblendvb
+      {TTI::SK_Select, MVT::v32i8, {1, 1, 1, 1}},  // vpblendvb
+
+      {TTI::SK_Splice, MVT::v8i32, {2, 2, 2, 2}},  // vperm2i128 + vpalignr
+      {TTI::SK_Splice, MVT::v8f32, {2, 2, 2, 2}},  // vperm2i128 + vpalignr
+      {TTI::SK_Splice, MVT::v16i16, {2, 2, 2, 2}}, // vperm2i128 + vpalignr
+      {TTI::SK_Splice, MVT::v16f16, {2, 2, 2, 2}}, // vperm2i128 + vpalignr
+      {TTI::SK_Splice, MVT::v32i8, {2, 2, 2, 2}},  // vperm2i128 + vpalignr
+
+      {TTI::SK_PermuteSingleSrc, MVT::v4f64, {1, 1, 1, 1}}, // vpermpd
+      {TTI::SK_PermuteSingleSrc, MVT::v8f32, {1, 1, 1, 1}}, // vpermps
+      {TTI::SK_PermuteSingleSrc, MVT::v4i64, {1, 1, 1, 1}}, // vpermq
+      {TTI::SK_PermuteSingleSrc, MVT::v8i32, {1, 1, 1, 1}}, // vpermd
+      {TTI::SK_PermuteSingleSrc, MVT::v16i16, {4, 4, 4, 4}},
+      {TTI::SK_PermuteSingleSrc, MVT::v16f16, {4, 4, 4, 4}},
+      {TTI::SK_PermuteSingleSrc, MVT::v32i8, {4, 4, 4, 4}},
+
+      {TTI::SK_PermuteTwoSrc, MVT::v4f64, {3, 3, 3, 3}}, // 2*vpermpd + vblendpd
+      {TTI::SK_PermuteTwoSrc, MVT::v8f32, {3, 3, 3, 3}}, // 2*vpermps + vblendps
+      {TTI::SK_PermuteTwoSrc, MVT::v4i64, {3, 3, 3, 3}}, // 2*vpermq + vpblendd
+      {TTI::SK_PermuteTwoSrc, MVT::v8i32, {3, 3, 3, 3}}, // 2*vpermd + vpblendd
+      {TTI::SK_PermuteTwoSrc, MVT::v16i16, {7, 7, 7, 7}},
+      {TTI::SK_PermuteTwoSrc, MVT::v16f16, {7, 7, 7, 7}},
+      {TTI::SK_PermuteTwoSrc, MVT::v32i8, {7, 7, 7, 7}},
   };
 
   if (ST->hasAVX2())
@@ -2088,62 +2088,100 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
         return LT.first * *KindCost;
 
   static const CostKindTblEntry AVX1ShuffleTbl[] = {
-      {TTI::SK_Broadcast, MVT::v4f64,  {2,3,2,3}}, // vperm2f128 + vpermilpd
-      {TTI::SK_Broadcast, MVT::v8f32,  {2,3,2,3}}, // vperm2f128 + vpermilps
-      {TTI::SK_Broadcast, MVT::v4i64,  {2,3,2,3}}, // vperm2f128 + vpermilpd
-      {TTI::SK_Broadcast, MVT::v8i32,  {2,3,2,3}}, // vperm2f128 + vpermilps
-      {TTI::SK_Broadcast, MVT::v16i16, {2,3,3,4}}, // vpshuflw + vpshufd + vinsertf128
-      {TTI::SK_Broadcast, MVT::v16f16, {2,3,3,4}}, // vpshuflw + vpshufd + vinsertf128
-      {TTI::SK_Broadcast, MVT::v32i8,  {3,4,3,6}}, // vpshufb + vinsertf128
-
-      {TTI::SK_Reverse, MVT::v4f64,  {2,6,2,2}}, // vperm2f128 + vpermilpd
-      {TTI::SK_Reverse, MVT::v8f32,  {2,7,2,4}}, // vperm2f128 + vpermilps
-      {TTI::SK_Reverse, MVT::v4i64,  {2,6,2,2}}, // vperm2f128 + vpermilpd
-      {TTI::SK_Reverse, MVT::v8i32,  {2,7,2,4}}, // vperm2f128 + vpermilps
-      {TTI::SK_Reverse, MVT::v16i16, {2,9,5,5}}, // vextractf128 + 2*pshufb
-                                                 // + vinsertf128
-      {TTI::SK_Reverse, MVT::v16f16, {2,9,5,5}}, // vextractf128 + 2*pshufb
-                                                 // + vinsertf128
-      {TTI::SK_Reverse, MVT::v32i8,  {2,9,5,5}}, // vextractf128 + 2*pshufb
-                                                 // + vinsertf128
-
-      {TTI::SK_Select, MVT::v4i64,  {1,1,1,1}}, // vblendpd
-      {TTI::SK_Select, MVT::v4f64,  {1,1,1,1}}, // vblendpd
-      {TTI::SK_Select, MVT::v8i32,  {1,1,1,1}}, // vblendps
-      {TTI::SK_Select, MVT::v8f32,  {1,1,1,1}}, // vblendps
-      {TTI::SK_Select, MVT::v16i16, {3,3,3,3}}, // vpand + vpandn + vpor
-      {TTI::SK_Select, MVT::v16f16, {3,3,3,3}}, // vpand + vpandn + vpor
-      {TTI::SK_Select, MVT::v32i8,  {3,3,3,3}}, // vpand + vpandn + vpor
-
-      {TTI::SK_Splice, MVT::v4i64,  {2,2,2,2}}, // vperm2f128 + shufpd
-      {TTI::SK_Splice, MVT::v4f64,  {2,2,2,2}}, // vperm2f128 + shufpd
-      {TTI::SK_Splice, MVT::v8i32,  {4,4,4,4}}, // 2*vperm2f128 + 2*vshufps
-      {TTI::SK_Splice, MVT::v8f32,  {4,4,4,4}}, // 2*vperm2f128 + 2*vshufps
-      {TTI::SK_Splice, MVT::v16i16, {5,5,5,5}}, // 2*vperm2f128 + 2*vpalignr + vinsertf128
-      {TTI::SK_Splice, MVT::v16f16, {5,5,5,5}}, // 2*vperm2f128 + 2*vpalignr + vinsertf128
-      {TTI::SK_Splice, MVT::v32i8,  {5,5,5,5}}, // 2*vperm2f128 + 2*vpalignr + vinsertf128
-
-      {TTI::SK_PermuteSingleSrc, MVT::v4f64, {2,2,2,2}}, // vperm2f128 + vshufpd
-      {TTI::SK_PermuteSingleSrc, MVT::v4i64, {2,2,2,2}}, // vperm2f128 + vshufpd
-      {TTI::SK_PermuteSingleSrc, MVT::v8f32, {4,4,4,4}}, // 2*vperm2f128 + 2*vshufps
-      {TTI::SK_PermuteSingleSrc, MVT::v8i32, {4,4,4,4}}, // 2*vperm2f128 + 2*vshufps
-      {TTI::SK_PermuteSingleSrc, MVT::v16i16,{8,8,8,8}}, // vextractf128 + 4*pshufb
-                                                         // + 2*por + vinsertf128
-      {TTI::SK_PermuteSingleSrc, MVT::v16f16,{8,8,8,8}}, // vextractf128 + 4*pshufb
-                                                         // + 2*por + vinsertf128
-      {TTI::SK_PermuteSingleSrc, MVT::v32i8, {8,8,8,8}}, // vextractf128 + 4*pshufb
-                                                         // + 2*por + vinsertf128
-
-      {TTI::SK_PermuteTwoSrc, MVT::v4f64, {3,3,3,3}},   // 2*vperm2f128 + vshufpd
-      {TTI::SK_PermuteTwoSrc, MVT::v4i64, {3,3,3,3}},   // 2*vperm2f128 + vshufpd
-      {TTI::SK_PermuteTwoSrc, MVT::v8f32, {4,4,4,4}},   // 2*vperm2f128 + 2*vshufps
-      {TTI::SK_PermuteTwoSrc, MVT::v8i32, {4,4,4,4}},   // 2*vperm2f128 + 2*vshufps
-      {TTI::SK_PermuteTwoSrc, MVT::v16i16,{15,15,15,15}}, // 2*vextractf128 + 8*pshufb
-                                                          // + 4*por + vinsertf128
-      {TTI::SK_PermuteTwoSrc, MVT::v16f16,{15,15,15,15}}, // 2*vextractf128 + 8*pshufb
-                                                          // + 4*por + vinsertf128
-      {TTI::SK_PermuteTwoSrc, MVT::v32i8, {15,15,15,15}}, // 2*vextractf128 + 8*pshufb
-                                                         // + 4*por + vinsertf128
+      {TTI::SK_Broadcast, MVT::v4f64, {2, 3, 2, 3}}, // vperm2f128 + vpermilpd
+      {TTI::SK_Broadcast, MVT::v8f32, {2, 3, 2, 3}}, // vperm2f128 + vpermilps
+      {TTI::SK_Broadcast, MVT::v4i64, {2, 3, 2, 3}}, // vperm2f128 + vpermilpd
+      {TTI::SK_Broadcast, MVT::v8i32, {2, 3, 2, 3}}, // vperm2f128 + vpermilps
+      {TTI::SK_Broadcast,
+       MVT::v16i16,
+       {2, 3, 3, 4}}, // vpshuflw + vpshufd + vinsertf128
+      {TTI::SK_Broadcast,
+       MVT::v16f16,
+       {2, 3, 3, 4}}, // vpshuflw + vpshufd + vinsertf128
+      {TTI::SK_Broadcast, MVT::v32i8, {3, 4, 3, 6}}, // vpshufb + vinsertf128
+
+      {TTI::SK_Reverse, MVT::v4f64, {2, 6, 2, 2}},  // vperm2f128 + vpermilpd
+      {TTI::SK_Reverse, MVT::v8f32, {2, 7, 2, 4}},  // vperm2f128 + vpermilps
+      {TTI::SK_Reverse, MVT::v4i64, {2, 6, 2, 2}},  // vperm2f128 + vpermilpd
+      {TTI::SK_Reverse, MVT::v8i32, {2, 7, 2, 4}},  // vperm2f128 + vpermilps
+      {TTI::SK_Reverse, MVT::v16i16, {2, 9, 5, 5}}, // vextractf128 + 2*pshufb
+                                                    // + vinsertf128
+      {TTI::SK_Reverse, MVT::v16f16, {2, 9, 5, 5}}, // vextractf128 + 2*pshufb
+                                                    // + vinsertf128
+      {TTI::SK_Reverse, MVT::v32i8, {2, 9, 5, 5}},  // vextractf128 + 2*pshufb
+                                                    // + vinsertf128
+
+      {TTI::SK_Select, MVT::v4i64, {1, 1, 1, 1}},  // vblendpd
+      {TTI::SK_Select, MVT::v4f64, {1, 1, 1, 1}},  // vblendpd
+      {TTI::SK_Select, MVT::v8i32, {1, 1, 1, 1}},  // vblendps
+      {TTI::SK_Select, MVT::v8f32, {1, 1, 1, 1}},  // vblendps
+      {TTI::SK_Select, MVT::v16i16, {3, 3, 3, 3}}, // vpand + vpandn + vpor
+      {TTI::SK_Select, MVT::v16f16, {3, 3, 3, 3}}, // vpand + vpandn + vpor
+      {TTI::SK_Select, MVT::v32i8, {3, 3, 3, 3}},  // vpand + vpandn + vpor
+
+      {TTI::SK_Splice, MVT::v4i64, {2, 2, 2, 2}}, // vperm2f128 + shufpd
+      {TTI::SK_Splice, MVT::v4f64, {2, 2, 2, 2}}, // vperm2f128 + shufpd
+      {TTI::SK_Splice, MVT::v8i32, {4, 4, 4, 4}}, // 2*vperm2f128 + 2*vshufps
+      {TTI::SK_Splice, MVT::v8f32, {4, 4, 4, 4}}, // 2*vperm2f128 + 2*vshufps
+      {TTI::SK_Splice,
+       MVT::v16i16,
+       {5, 5, 5, 5}}, // 2*vperm2f128 + 2*vpalignr + vinsertf128
+      {TTI::SK_Splice,
+       MVT::v16f16,
+       {5, 5, 5, 5}}, // 2*vperm2f128 + 2*vpalignr + vinsertf128
+      {TTI::SK_Splice,
+       MVT::v32i8,
+       {5, 5, 5, 5}}, // 2*vperm2f128 + 2*vpalignr + vinsertf128
+
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v4f64,
+       {2, 2, 2, 2}}, // vperm2f128 + vshufpd
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v4i64,
+       {2, 2, 2, 2}}, // vperm2f128 + vshufpd
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v8f32,
+       {4, 4, 4, 4}}, // 2*vperm2f128 + 2*vshufps
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v8i32,
+       {4, 4, 4, 4}}, // 2*vperm2f128 + 2*vshufps
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v16i16,
+       {8, 8, 8, 8}}, // vextractf128 + 4*pshufb
+                      // + 2*por + vinsertf128
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v16f16,
+       {8, 8, 8, 8}}, // vextractf128 + 4*pshufb
+                      // + 2*por + vinsertf128
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v32i8,
+       {8, 8, 8, 8}}, // vextractf128 + 4*pshufb
+                      // + 2*por + vinsertf128
+
+      {TTI::SK_PermuteTwoSrc,
+       MVT::v4f64,
+       {3, 3, 3, 3}}, // 2*vperm2f128 + vshufpd
+      {TTI::SK_PermuteTwoSrc,
+       MVT::v4i64,
+       {3, 3, 3, 3}}, // 2*vperm2f128 + vshufpd
+      {TTI::SK_PermuteTwoSrc,
+       MVT::v8f32,
+       {4, 4, 4, 4}}, // 2*vperm2f128 + 2*vshufps
+      {TTI::SK_PermuteTwoSrc,
+       MVT::v8i32,
+       {4, 4, 4, 4}}, // 2*vperm2f128 + 2*vshufps
+      {TTI::SK_PermuteTwoSrc,
+       MVT::v16i16,
+       {15, 15, 15, 15}}, // 2*vextractf128 + 8*pshufb
+                          // + 4*por + vinsertf128
+      {TTI::SK_PermuteTwoSrc,
+       MVT::v16f16,
+       {15, 15, 15, 15}}, // 2*vextractf128 + 8*pshufb
+                          // + 4*por + vinsertf128
+      {TTI::SK_PermuteTwoSrc,
+       MVT::v32i8,
+       {15, 15, 15, 15}}, // 2*vextractf128 + 8*pshufb
+                          // + 4*por + vinsertf128
   };
 
   if (ST->hasAVX())
@@ -2212,8 +2250,10 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
       {TTI::SK_Reverse, MVT::v4i32, {1, 1, 1, 1}}, // pshufd
       {TTI::SK_Reverse, MVT::v8i16, {2, 3, 3, 3}}, // pshuflw + pshufhw + pshufd
       {TTI::SK_Reverse, MVT::v8f16, {2, 3, 3, 3}}, // pshuflw + pshufhw + pshufd
-      {TTI::SK_Reverse, MVT::v16i8, {5, 6,11,11}}, // 2*pshuflw + 2*pshufhw
-                                                   // + 2*pshufd + 2*unpck + packus
+      {TTI::SK_Reverse,
+       MVT::v16i8,
+       {5, 6, 11, 11}}, // 2*pshuflw + 2*pshufhw
+                        // + 2*pshufd + 2*unpck + packus
 
       {TTI::SK_Select, MVT::v2i64, {1, 1, 1, 1}}, // movsd
       {TTI::SK_Select, MVT::v2f64, {1, 1, 1, 1}}, // movsd
@@ -2232,16 +2272,24 @@ InstructionCost X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
       {TTI::SK_PermuteSingleSrc, MVT::v2f64, {1, 1, 1, 1}}, // shufpd
       {TTI::SK_PermuteSingleSrc, MVT::v2i64, {1, 1, 1, 1}}, // pshufd
       {TTI::SK_PermuteSingleSrc, MVT::v4i32, {1, 1, 1, 1}}, // pshufd
-      {TTI::SK_PermuteSingleSrc, MVT::v8i16, {3, 5, 5, 5}}, // 2*pshuflw + 2*pshufhw
-                                                            // + pshufd/unpck
-      {TTI::SK_PermuteSingleSrc, MVT::v8f16, {3, 5, 5, 5}}, // 2*pshuflw + 2*pshufhw
-                                                            // + pshufd/unpck
-      {TTI::SK_PermuteSingleSrc, MVT::v16i8, {8, 10, 10, 10}}, // 2*pshuflw + 2*pshufhw
-                                                               // + 2*pshufd + 2*unpck + 2*packus
-
-      {TTI::SK_PermuteTwoSrc, MVT::v2f64, {1, 1, 1, 1}},     // shufpd
-      {TTI::SK_PermuteTwoSrc, MVT::v2i64, {1, 1, 1, 1}},     // shufpd
-      {TTI::SK_PermuteTwoSrc, MVT::v4i32, {2, 2, 2, 2}},     // 2*{unpck,movsd,pshufd}
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v8i16,
+       {3, 5, 5, 5}}, // 2*pshuflw + 2*pshufhw
+                      // + pshufd/unpck
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v8f16,
+       {3, 5, 5, 5}}, // 2*pshuflw + 2*pshufhw
+                      // + pshufd/unpck
+      {TTI::SK_PermuteSingleSrc,
+       MVT::v16i8,
+       {8, 10, 10, 10}}, // 2*pshuflw + 2*pshufhw
+                         // + 2*pshufd + 2*unpck + 2*packus
+
+      {TTI::SK_PermuteTwoSrc, MVT::v2f64, {1, 1, 1, 1}}, // shufpd
+      {TTI::SK_PermuteTwoSrc, MVT::v2i64, {1, 1, 1, 1}}, // shufpd
+      {TTI::SK_PermuteTwoSrc,
+       MVT::v4i32,
+       {2, 2, 2, 2}}, // 2*{unpck,movsd,pshufd}
       {TTI::SK_PermuteTwoSrc, MVT::v8i16, {6, 8, 8, 8}},     // blend+permute
       {TTI::SK_PermuteTwoSrc, MVT::v8f16, {6, 8, 8, 8}},     // blend+permute
       {TTI::SK_PermuteTwoSrc, MVT::v16i8, {11, 13, 13, 13}}, // blend+permute

@RKSimon RKSimon merged commit 3820206 into llvm:main Jul 26, 2025
8 of 9 checks passed
@RKSimon RKSimon deleted the x86-costs-sk_reverse branch July 26, 2025 17:22
mahesh-attarde pushed a commit to mahesh-attarde/llvm-project that referenced this pull request Jul 28, 2025
@RKSimon @mahesh-attarde
[CostModel][X86] Update SK_Reverse based on cost kinds (llvm#150650)
af1d3af 
When these were converted to CostKindTblEntry the throughput was mainly
copied to all cost kinds

Regenerated with my check_cost_tables.py helper script
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
backend:X86 llvm:analysis Includes value tracking, cost tables and constant folding
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@RKSimon @llvmbot