[llvm] 5a7dfb4 - [CostModel][X86] Attempt to match v4f32 shuffles that map to MOVSS/INSERTPS instruction
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Tue Jan 7 03:32:02 PST 2025
Author: Simon Pilgrim
Date: 2025-01-07T11:31:44Z
New Revision: 5a7dfb46598ebe21318312eca5f407a1575f87e0
URL: https://github.com/llvm/llvm-project/commit/5a7dfb46598ebe21318312eca5f407a1575f87e0
DIFF: https://github.com/llvm/llvm-project/commit/5a7dfb46598ebe21318312eca5f407a1575f87e0.diff
LOG: [CostModel][X86] Attempt to match v4f32 shuffles that map to MOVSS/INSERTPS instruction
improveShuffleKindFromMask matches this as a SK_InsertSubvector of a v1f32 (which legalises to f32) into a v4f32 base vector, making it easy to recognise. MOVSS is limited to index0.
Added:
Modified:
llvm/lib/Target/X86/X86TargetTransformInfo.cpp
llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll
llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll
llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll
llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll
llvm/test/Transforms/PhaseOrdering/X86/hadd.ll
llvm/test/Transforms/PhaseOrdering/X86/hsub.ll
llvm/test/Transforms/VectorCombine/X86/extract-binop-inseltpoison.ll
llvm/test/Transforms/VectorCombine/X86/extract-binop.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index 7a7554cdb99088..d26dec8f7b9504 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -1650,6 +1650,13 @@ InstructionCost X86TTIImpl::getShuffleCost(
return MatchingTypes ? TTI::TCC_Free : SubLT.first;
}
+ // Attempt to match MOVSS (Idx == 0) or INSERTPS pattern. This will have
+ // been matched by improveShuffleKindFromMask as a SK_InsertSubvector of
+ // v1f32 (legalised to f32) into a v4f32.
+ if (LT.first == 1 && LT.second == MVT::v4f32 && SubLT.first == 1 &&
+ SubLT.second == MVT::f32 && (Index == 0 || ST->hasSSE41()))
+ return 1;
+
// If the insertion isn't aligned, treat it like a 2-op shuffle.
Kind = TTI::SK_PermuteTwoSrc;
}
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll
index 027af628ea3228..b672df58685219 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll
@@ -124,7 +124,7 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, 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 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll
index f9f045f3a172bc..fc8c0cd0a3998b 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll
@@ -124,7 +124,7 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, 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 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll
index 76690afecabdd4..b48b6205a2550f 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll
@@ -124,7 +124,7 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, 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 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll
index 034ec0acf79d9f..efa0f2eb8dc941 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll
@@ -124,7 +124,7 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, 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 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
diff --git a/llvm/test/Transforms/PhaseOrdering/X86/hadd.ll b/llvm/test/Transforms/PhaseOrdering/X86/hadd.ll
index 3c72d385dcf48b..a4aea02a335117 100644
--- a/llvm/test/Transforms/PhaseOrdering/X86/hadd.ll
+++ b/llvm/test/Transforms/PhaseOrdering/X86/hadd.ll
@@ -566,22 +566,16 @@ define <4 x float> @add_v4f32_012u(<4 x float> %a, <4 x float> %b) {
; SSE2-NEXT: ret <4 x float> [[RESULT1]]
;
; SSE4-LABEL: @add_v4f32_012u(
-; SSE4-NEXT: [[SHIFT:%.*]] = shufflevector <4 x float> [[B:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 poison, i32 poison, i32 poison>
-; SSE4-NEXT: [[TMP1:%.*]] = fadd <4 x float> [[B]], [[SHIFT]]
-; SSE4-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 2, i32 poison, i32 poison>
-; SSE4-NEXT: [[TMP3:%.*]] = shufflevector <4 x float> [[A]], <4 x float> poison, <4 x i32> <i32 0, i32 3, i32 poison, i32 poison>
+; SSE4-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x i32> <i32 1, i32 2, i32 4, i32 poison>
+; SSE4-NEXT: [[TMP3:%.*]] = shufflevector <4 x float> [[A]], <4 x float> [[B]], <4 x i32> <i32 0, i32 3, i32 5, i32 poison>
; SSE4-NEXT: [[TMP4:%.*]] = fadd <4 x float> [[TMP2]], [[TMP3]]
-; SSE4-NEXT: [[RESULT:%.*]] = shufflevector <4 x float> [[TMP4]], <4 x float> [[TMP1]], <4 x i32> <i32 0, i32 1, i32 4, i32 poison>
-; SSE4-NEXT: ret <4 x float> [[RESULT]]
+; SSE4-NEXT: ret <4 x float> [[TMP4]]
;
; AVX2-LABEL: @add_v4f32_012u(
-; AVX2-NEXT: [[SHIFT:%.*]] = shufflevector <4 x float> [[B:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 poison, i32 poison, i32 poison>
-; AVX2-NEXT: [[TMP1:%.*]] = fadd <4 x float> [[B]], [[SHIFT]]
-; AVX2-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 2, i32 poison, i32 poison>
-; AVX2-NEXT: [[TMP3:%.*]] = shufflevector <4 x float> [[A]], <4 x float> poison, <4 x i32> <i32 0, i32 3, i32 poison, i32 poison>
+; AVX2-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x i32> <i32 1, i32 2, i32 4, i32 poison>
+; AVX2-NEXT: [[TMP3:%.*]] = shufflevector <4 x float> [[A]], <4 x float> [[B]], <4 x i32> <i32 0, i32 3, i32 5, i32 poison>
; AVX2-NEXT: [[TMP4:%.*]] = fadd <4 x float> [[TMP2]], [[TMP3]]
-; AVX2-NEXT: [[RESULT:%.*]] = shufflevector <4 x float> [[TMP4]], <4 x float> [[TMP1]], <4 x i32> <i32 0, i32 1, i32 4, i32 poison>
-; AVX2-NEXT: ret <4 x float> [[RESULT]]
+; AVX2-NEXT: ret <4 x float> [[TMP4]]
;
; AVX512-LABEL: @add_v4f32_012u(
; AVX512-NEXT: [[TMP1:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x i32> <i32 0, i32 2, i32 4, i32 poison>
diff --git a/llvm/test/Transforms/PhaseOrdering/X86/hsub.ll b/llvm/test/Transforms/PhaseOrdering/X86/hsub.ll
index cf1c948ac6af58..bcb316a4a73ea6 100644
--- a/llvm/test/Transforms/PhaseOrdering/X86/hsub.ll
+++ b/llvm/test/Transforms/PhaseOrdering/X86/hsub.ll
@@ -566,22 +566,16 @@ define <4 x float> @sub_v4f32_012u(<4 x float> %a, <4 x float> %b) {
; SSE2-NEXT: ret <4 x float> [[RESULT1]]
;
; SSE4-LABEL: @sub_v4f32_012u(
-; SSE4-NEXT: [[SHIFT:%.*]] = shufflevector <4 x float> [[B:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 poison, i32 poison, i32 poison>
-; SSE4-NEXT: [[TMP1:%.*]] = fsub <4 x float> [[B]], [[SHIFT]]
-; SSE4-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> poison, <4 x i32> <i32 0, i32 2, i32 poison, i32 poison>
-; SSE4-NEXT: [[TMP3:%.*]] = shufflevector <4 x float> [[A]], <4 x float> poison, <4 x i32> <i32 1, i32 3, i32 poison, i32 poison>
+; SSE4-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x i32> <i32 0, i32 2, i32 4, i32 poison>
+; SSE4-NEXT: [[TMP3:%.*]] = shufflevector <4 x float> [[A]], <4 x float> [[B]], <4 x i32> <i32 1, i32 3, i32 5, i32 poison>
; SSE4-NEXT: [[TMP4:%.*]] = fsub <4 x float> [[TMP2]], [[TMP3]]
-; SSE4-NEXT: [[RESULT:%.*]] = shufflevector <4 x float> [[TMP4]], <4 x float> [[TMP1]], <4 x i32> <i32 0, i32 1, i32 4, i32 poison>
-; SSE4-NEXT: ret <4 x float> [[RESULT]]
+; SSE4-NEXT: ret <4 x float> [[TMP4]]
;
; AVX2-LABEL: @sub_v4f32_012u(
-; AVX2-NEXT: [[SHIFT:%.*]] = shufflevector <4 x float> [[B:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 poison, i32 poison, i32 poison>
-; AVX2-NEXT: [[TMP1:%.*]] = fsub <4 x float> [[B]], [[SHIFT]]
-; AVX2-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> poison, <4 x i32> <i32 0, i32 2, i32 poison, i32 poison>
-; AVX2-NEXT: [[TMP3:%.*]] = shufflevector <4 x float> [[A]], <4 x float> poison, <4 x i32> <i32 1, i32 3, i32 poison, i32 poison>
+; AVX2-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x i32> <i32 0, i32 2, i32 4, i32 poison>
+; AVX2-NEXT: [[TMP3:%.*]] = shufflevector <4 x float> [[A]], <4 x float> [[B]], <4 x i32> <i32 1, i32 3, i32 5, i32 poison>
; AVX2-NEXT: [[TMP4:%.*]] = fsub <4 x float> [[TMP2]], [[TMP3]]
-; AVX2-NEXT: [[RESULT:%.*]] = shufflevector <4 x float> [[TMP4]], <4 x float> [[TMP1]], <4 x i32> <i32 0, i32 1, i32 4, i32 poison>
-; AVX2-NEXT: ret <4 x float> [[RESULT]]
+; AVX2-NEXT: ret <4 x float> [[TMP4]]
;
; AVX512-LABEL: @sub_v4f32_012u(
; AVX512-NEXT: [[TMP1:%.*]] = shufflevector <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x i32> <i32 0, i32 2, i32 4, i32 poison>
diff --git a/llvm/test/Transforms/VectorCombine/X86/extract-binop-inseltpoison.ll b/llvm/test/Transforms/VectorCombine/X86/extract-binop-inseltpoison.ll
index 6ef18e66d4211f..f3b7f7b72ee425 100644
--- a/llvm/test/Transforms/VectorCombine/X86/extract-binop-inseltpoison.ll
+++ b/llvm/test/Transforms/VectorCombine/X86/extract-binop-inseltpoison.ll
@@ -465,28 +465,13 @@ define <4 x float> @ins_bo_ext_ext_uses(<4 x float> %a, <4 x float> %b) {
}
define <4 x float> @PR34724(<4 x float> %a, <4 x float> %b) {
-; SSE-LABEL: @PR34724(
-; SSE-NEXT: [[A0:%.*]] = extractelement <4 x float> [[A:%.*]], i32 0
-; SSE-NEXT: [[A1:%.*]] = extractelement <4 x float> [[A]], i32 1
-; SSE-NEXT: [[SHIFT2:%.*]] = shufflevector <4 x float> [[A]], <4 x float> [[B1:%.*]], <4 x i32> <i32 poison, i32 2, i32 4, i32 6>
-; SSE-NEXT: [[B:%.*]] = shufflevector <4 x float> [[A]], <4 x float> [[B1]], <4 x i32> <i32 poison, i32 3, i32 5, i32 7>
-; SSE-NEXT: [[TMP3:%.*]] = fadd <4 x float> [[SHIFT2]], [[B]]
-; SSE-NEXT: ret <4 x float> [[TMP3]]
-;
-; AVX-LABEL: @PR34724(
-; AVX-NEXT: [[A0:%.*]] = extractelement <4 x float> [[A:%.*]], i32 0
-; AVX-NEXT: [[A1:%.*]] = extractelement <4 x float> [[A]], i32 1
-; AVX-NEXT: [[SHIFT:%.*]] = shufflevector <4 x float> [[A]], <4 x float> poison, <4 x i32> <i32 poison, i32 poison, i32 3, i32 poison>
-; AVX-NEXT: [[TMP1:%.*]] = fadd <4 x float> [[A]], [[SHIFT]]
-; AVX-NEXT: [[SHIFT1:%.*]] = shufflevector <4 x float> [[B:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT: [[TMP2:%.*]] = fadd <4 x float> [[B]], [[SHIFT1]]
-; AVX-NEXT: [[B01:%.*]] = extractelement <4 x float> [[TMP2]], i32 0
-; AVX-NEXT: [[SHIFT2:%.*]] = shufflevector <4 x float> [[B]], <4 x float> poison, <4 x i32> <i32 poison, i32 poison, i32 poison, i32 2>
-; AVX-NEXT: [[TMP3:%.*]] = fadd <4 x float> [[SHIFT2]], [[B]]
-; AVX-NEXT: [[V1:%.*]] = shufflevector <4 x float> [[TMP1]], <4 x float> poison, <4 x i32> <i32 poison, i32 2, i32 poison, i32 poison>
-; AVX-NEXT: [[V2:%.*]] = insertelement <4 x float> [[V1]], float [[B01]], i32 2
-; AVX-NEXT: [[V3:%.*]] = shufflevector <4 x float> [[V2]], <4 x float> [[TMP3]], <4 x i32> <i32 0, i32 1, i32 2, i32 7>
-; AVX-NEXT: ret <4 x float> [[V3]]
+; CHECK-LABEL: @PR34724(
+; CHECK-NEXT: [[A0:%.*]] = extractelement <4 x float> [[A:%.*]], i32 0
+; CHECK-NEXT: [[A1:%.*]] = extractelement <4 x float> [[A]], i32 1
+; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <4 x float> [[A]], <4 x float> [[B:%.*]], <4 x i32> <i32 poison, i32 2, i32 4, i32 6>
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x float> [[A]], <4 x float> [[B]], <4 x i32> <i32 poison, i32 3, i32 5, i32 7>
+; CHECK-NEXT: [[V3:%.*]] = fadd <4 x float> [[TMP1]], [[TMP2]]
+; CHECK-NEXT: ret <4 x float> [[V3]]
;
%a0 = extractelement <4 x float> %a, i32 0
%a1 = extractelement <4 x float> %a, i32 1
diff --git a/llvm/test/Transforms/VectorCombine/X86/extract-binop.ll b/llvm/test/Transforms/VectorCombine/X86/extract-binop.ll
index 307fbf711cdc2d..c125b73fccddf9 100644
--- a/llvm/test/Transforms/VectorCombine/X86/extract-binop.ll
+++ b/llvm/test/Transforms/VectorCombine/X86/extract-binop.ll
@@ -465,34 +465,19 @@ define <4 x float> @ins_bo_ext_ext_uses(<4 x float> %a, <4 x float> %b) {
}
define <4 x float> @PR34724(<4 x float> %a, <4 x float> %b) {
-; SSE-LABEL: @PR34724(
-; SSE-NEXT: [[A0:%.*]] = extractelement <4 x float> [[A:%.*]], i32 0
-; SSE-NEXT: [[A1:%.*]] = extractelement <4 x float> [[A]], i32 1
-; SSE-NEXT: [[SHIFT:%.*]] = shufflevector <4 x float> [[A]], <4 x float> poison, <4 x i32> <i32 poison, i32 poison, i32 3, i32 poison>
-; SSE-NEXT: [[TMP1:%.*]] = fadd <4 x float> [[A]], [[SHIFT]]
-; SSE-NEXT: [[SHIFT1:%.*]] = shufflevector <4 x float> [[B:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 poison, i32 poison, i32 poison>
-; SSE-NEXT: [[TMP2:%.*]] = fadd <4 x float> [[B]], [[SHIFT1]]
-; SSE-NEXT: [[SHIFT2:%.*]] = shufflevector <4 x float> [[B]], <4 x float> poison, <4 x i32> <i32 poison, i32 poison, i32 poison, i32 2>
-; SSE-NEXT: [[TMP3:%.*]] = fadd <4 x float> [[SHIFT2]], [[B]]
-; SSE-NEXT: [[V1:%.*]] = shufflevector <4 x float> [[TMP1]], <4 x float> undef, <4 x i32> <i32 4, i32 2, i32 6, i32 7>
-; SSE-NEXT: [[V2:%.*]] = shufflevector <4 x float> [[V1]], <4 x float> [[TMP2]], <4 x i32> <i32 0, i32 1, i32 4, i32 3>
-; SSE-NEXT: [[V3:%.*]] = shufflevector <4 x float> [[V2]], <4 x float> [[TMP3]], <4 x i32> <i32 0, i32 1, i32 2, i32 7>
-; SSE-NEXT: ret <4 x float> [[V3]]
-;
-; AVX-LABEL: @PR34724(
-; AVX-NEXT: [[A0:%.*]] = extractelement <4 x float> [[A:%.*]], i32 0
-; AVX-NEXT: [[A1:%.*]] = extractelement <4 x float> [[A]], i32 1
-; AVX-NEXT: [[SHIFT:%.*]] = shufflevector <4 x float> [[A]], <4 x float> poison, <4 x i32> <i32 poison, i32 poison, i32 3, i32 poison>
-; AVX-NEXT: [[TMP1:%.*]] = fadd <4 x float> [[A]], [[SHIFT]]
-; AVX-NEXT: [[SHIFT1:%.*]] = shufflevector <4 x float> [[B:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 poison, i32 poison, i32 poison>
-; AVX-NEXT: [[TMP2:%.*]] = fadd <4 x float> [[B]], [[SHIFT1]]
-; AVX-NEXT: [[B01:%.*]] = extractelement <4 x float> [[TMP2]], i32 0
-; AVX-NEXT: [[SHIFT2:%.*]] = shufflevector <4 x float> [[B]], <4 x float> poison, <4 x i32> <i32 poison, i32 poison, i32 poison, i32 2>
-; AVX-NEXT: [[TMP3:%.*]] = fadd <4 x float> [[SHIFT2]], [[B]]
-; AVX-NEXT: [[V1:%.*]] = shufflevector <4 x float> [[TMP1]], <4 x float> undef, <4 x i32> <i32 4, i32 2, i32 6, i32 7>
-; AVX-NEXT: [[V2:%.*]] = insertelement <4 x float> [[V1]], float [[B01]], i32 2
-; AVX-NEXT: [[V3:%.*]] = shufflevector <4 x float> [[V2]], <4 x float> [[TMP3]], <4 x i32> <i32 0, i32 1, i32 2, i32 7>
-; AVX-NEXT: ret <4 x float> [[V3]]
+; CHECK-LABEL: @PR34724(
+; CHECK-NEXT: [[A0:%.*]] = extractelement <4 x float> [[A:%.*]], i32 0
+; CHECK-NEXT: [[A1:%.*]] = extractelement <4 x float> [[A]], i32 1
+; CHECK-NEXT: [[SHIFT:%.*]] = shufflevector <4 x float> [[A]], <4 x float> poison, <4 x i32> <i32 poison, i32 poison, i32 3, i32 poison>
+; CHECK-NEXT: [[TMP1:%.*]] = fadd <4 x float> [[A]], [[SHIFT]]
+; CHECK-NEXT: [[SHIFT1:%.*]] = shufflevector <4 x float> [[B:%.*]], <4 x float> poison, <4 x i32> <i32 1, i32 poison, i32 poison, i32 poison>
+; CHECK-NEXT: [[TMP2:%.*]] = fadd <4 x float> [[B]], [[SHIFT1]]
+; CHECK-NEXT: [[SHIFT2:%.*]] = shufflevector <4 x float> [[B]], <4 x float> poison, <4 x i32> <i32 poison, i32 poison, i32 poison, i32 2>
+; CHECK-NEXT: [[TMP3:%.*]] = fadd <4 x float> [[SHIFT2]], [[B]]
+; CHECK-NEXT: [[V1:%.*]] = shufflevector <4 x float> [[TMP1]], <4 x float> undef, <4 x i32> <i32 4, i32 2, i32 6, i32 7>
+; CHECK-NEXT: [[V2:%.*]] = shufflevector <4 x float> [[V1]], <4 x float> [[TMP2]], <4 x i32> <i32 0, i32 1, i32 4, i32 3>
+; CHECK-NEXT: [[V3:%.*]] = shufflevector <4 x float> [[V2]], <4 x float> [[TMP3]], <4 x i32> <i32 0, i32 1, i32 2, i32 7>
+; CHECK-NEXT: ret <4 x float> [[V3]]
;
%a0 = extractelement <4 x float> %a, i32 0
%a1 = extractelement <4 x float> %a, i32 1
More information about the llvm-commits
mailing list