[llvm] [LV] Limits the splat operations be hoisted must not be defined by a recipe. (PR #117138)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-llvm-transforms

Author: Mel Chen (Mel-Chen)

<details>
<summary>Changes</summary>



---
Full diff: https://github.com/llvm/llvm-project/pull/117138.diff


4 Files Affected:

- (modified) llvm/lib/Transforms/Vectorize/VPlan.cpp (+1-1) 
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/extractvalue-no-scalarization-required.ll (+2-2) 
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-extractvalue.ll (+1-1) 
- (modified) llvm/test/Transforms/LoopVectorize/X86/epilog-vectorization-inductions.ll (+2-2) 


``````````diff
diff --git a/llvm/lib/Transforms/Vectorize/VPlan.cpp b/llvm/lib/Transforms/Vectorize/VPlan.cpp
index 8b1a4aeb88f81f..32996426c28490 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlan.cpp
@@ -266,7 +266,7 @@ Value *VPTransformState::get(VPValue *Def, bool NeedsScalar) {
     return Data.VPV2Vector[Def];
 
   auto GetBroadcastInstrs = [this, Def](Value *V) {
-    bool SafeToHoist = Def->isDefinedOutsideLoopRegions();
+    bool SafeToHoist = !Def->hasDefiningRecipe();
     if (VF.isScalar())
       return V;
     // Place the code for broadcasting invariant variables in the new preheader.
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/extractvalue-no-scalarization-required.ll b/llvm/test/Transforms/LoopVectorize/AArch64/extractvalue-no-scalarization-required.ll
index aa78113ebaa48c..b1c202eab9dd3d 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/extractvalue-no-scalarization-required.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/extractvalue-no-scalarization-required.ll
@@ -23,9 +23,9 @@
 ; Check that the extractvalue operands are actually free in vector code.
 
 ; FORCED:         [[E1:%.+]] = extractvalue { i64, i64 } %sv, 0
-; FORCED-NEXT:    [[E2:%.+]] = extractvalue { i64, i64 } %sv, 1
 ; FORCED-NEXT:    %broadcast.splatinsert = insertelement <2 x i64> poison, i64 [[E1]], i64 0
 ; FORCED-NEXT:    %broadcast.splat = shufflevector <2 x i64> %broadcast.splatinsert, <2 x i64> poison, <2 x i32> zeroinitializer
+; FORCED-NEXT:    [[E2:%.+]] = extractvalue { i64, i64 } %sv, 1
 ; FORCED-NEXT:    %broadcast.splatinsert1 = insertelement <2 x i64> poison, i64 [[E2]], i64 0
 ; FORCED-NEXT:    %broadcast.splat2 = shufflevector <2 x i64> %broadcast.splatinsert1, <2 x i64> poison, <2 x i32> zeroinitializer
 ; FORCED-NEXT:    [[ADD:%.+]] = add <2 x i64> %broadcast.splat, %broadcast.splat2
@@ -75,9 +75,9 @@ declare float @powf(float, float) readnone nounwind
 ; FORCED-LABEL: define void @test_getVectorCallCost
 
 ; FORCED:         [[E1:%.+]] = extractvalue { float, float } %sv, 0
-; FORCED-NEXT:    [[E2:%.+]] = extractvalue { float, float } %sv, 1
 ; FORCED-NEXT:    %broadcast.splatinsert = insertelement <2 x float> poison, float [[E1]], i64 0
 ; FORCED-NEXT:    %broadcast.splat = shufflevector <2 x float> %broadcast.splatinsert, <2 x float> poison, <2 x i32> zeroinitializer
+; FORCED-NEXT:    [[E2:%.+]] = extractvalue { float, float } %sv, 1
 ; FORCED-NEXT:    %broadcast.splatinsert1 = insertelement <2 x float> poison, float [[E2]], i64 0
 ; FORCED-NEXT:    %broadcast.splat2 = shufflevector <2 x float> %broadcast.splatinsert1, <2 x float> poison, <2 x i32> zeroinitializer
 
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-extractvalue.ll b/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-extractvalue.ll
index 7778f01c58dc34..fb5087db254b23 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-extractvalue.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-extractvalue.ll
@@ -5,9 +5,9 @@ target triple = "aarch64-unknown-linux-gnu"
 define void @widen_extractvalue(ptr %dst, {i64, i64} %sv) #0 {
 ; CHECK-LABEL: @widen_extractvalue(
 ; CHECK:        [[EXTRACT0:%.*]] = extractvalue { i64, i64 } [[SV:%.*]], 0
-; CHECK-NEXT:   [[EXTRACT1:%.*]] = extractvalue { i64, i64 } [[SV]], 1
 ; CHECK-NEXT:   [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[EXTRACT0]], i64 0
 ; CHECK-NEXT:   [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
+; CHECK-NEXT:   [[EXTRACT1:%.*]] = extractvalue { i64, i64 } [[SV]], 1
 ; CHECK-NEXT:   [[DOTSPLATINSERT1:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[EXTRACT1]], i64 0
 ; CHECK-NEXT:   [[DOTSPLAT2:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT1]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
 ; CHECK:        [[ADD:%.*]] = add <vscale x 2 x i64> [[DOTSPLAT]], [[DOTSPLAT2]]
diff --git a/llvm/test/Transforms/LoopVectorize/X86/epilog-vectorization-inductions.ll b/llvm/test/Transforms/LoopVectorize/X86/epilog-vectorization-inductions.ll
index e94bd841360256..7840a9dec794b3 100644
--- a/llvm/test/Transforms/LoopVectorize/X86/epilog-vectorization-inductions.ll
+++ b/llvm/test/Transforms/LoopVectorize/X86/epilog-vectorization-inductions.ll
@@ -137,6 +137,8 @@ define void @test_induction_step_needs_expansion(ptr noalias %j, ptr %k, i64 %l,
 ; CHECK-LABEL: @test_induction_step_needs_expansion(
 ; CHECK-NEXT:  iter.check:
 ; CHECK-NEXT:    [[TMP0:%.*]] = sub i16 0, [[OFF:%.*]]
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <16 x i16> poison, i16 [[TMP0]], i64 0
+; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <16 x i16> [[BROADCAST_SPLATINSERT]], <16 x i16> poison, <16 x i32> zeroinitializer
 ; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[L:%.*]], 8
 ; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
 ; CHECK:       vector.main.loop.iter.check:
@@ -145,8 +147,6 @@ define void @test_induction_step_needs_expansion(ptr noalias %j, ptr %k, i64 %l,
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[L]], 64
 ; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[L]], [[N_MOD_VF]]
-; CHECK-NEXT:    [[DOTSPLATINSERT2:%.*]] = insertelement <16 x i16> poison, i16 [[TMP0]], i64 0
-; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <16 x i16> [[DOTSPLATINSERT2]], <16 x i16> poison, <16 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP1:%.*]] = mul <16 x i16> splat (i16 16), [[TMP2]]
 ; CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i16> poison, i16 [[TMP0]], i64 0
 ; CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i16> [[DOTSPLATINSERT]], <16 x i16> poison, <16 x i32> zeroinitializer

``````````

</details>


https://github.com/llvm/llvm-project/pull/117138