[llvm] [SLP]Add runtime stride support for strided loads. (PR #81517)

Alexey Bataev via llvm-commits llvm-commits at lists.llvm.org
Tue Mar 5 06:35:40 PST 2024


https://github.com/alexey-bataev updated https://github.com/llvm/llvm-project/pull/81517

>From 72d585d454b69eebfd0426509ac68d49331e8d3e Mon Sep 17 00:00:00 2001
From: Alexey Bataev <a.bataev at outlook.com>
Date: Mon, 12 Feb 2024 19:05:25 +0000
Subject: [PATCH] =?UTF-8?q?[=F0=9D=98=80=F0=9D=97=BD=F0=9D=97=BF]=20initia?=
 =?UTF-8?q?l=20version?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Created using spr 1.3.5
---
 .../Transforms/Vectorize/SLPVectorizer.cpp    | 150 +++++++++++++++++-
 .../RISCV/strided-loads-vectorized.ll         |  98 ++----------
 2 files changed, 155 insertions(+), 93 deletions(-)

diff --git a/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp b/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
index c54d065cac6382..8fa32035d172d5 100644
--- a/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
+++ b/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
@@ -87,6 +87,7 @@
 #include "llvm/Transforms/Utils/InjectTLIMappings.h"
 #include "llvm/Transforms/Utils/Local.h"
 #include "llvm/Transforms/Utils/LoopUtils.h"
+#include "llvm/Transforms/Utils/ScalarEvolutionExpander.h"
 #include <algorithm>
 #include <cassert>
 #include <cstdint>
@@ -3890,6 +3891,122 @@ static bool isReverseOrder(ArrayRef<unsigned> Order) {
   });
 }
 
+/// Checks if the provided list of pointers \p Pointers represents the strided
+/// pointers for type ElemTy. If they are not, std::nullopt is returned.
+/// Otherwise, if \p Inst is not specified, just initialized optional value is
+/// returned to show that the pointers represent strided pointers. If \p Inst
+/// specified, the runtime stride is materialized before the given \p Inst.
+/// \returns std::nullopt if the pointers are not pointers with the runtime
+/// stride, nullptr or actual stride value, otherwise.
+static std::optional<Value *>
+calculateRtStride(ArrayRef<Value *> PointerOps, Type *ElemTy,
+                  const DataLayout &DL, ScalarEvolution &SE,
+                  SmallVectorImpl<unsigned> &SortedIndices,
+                  Instruction *Inst = nullptr) {
+  SmallVector<const SCEV *> SCEVs;
+  const SCEV *PtrSCEVA = nullptr;
+  const SCEV *PtrSCEVB = nullptr;
+  for (Value *Ptr : PointerOps) {
+    const SCEV *PtrSCEV = SE.getSCEV(Ptr);
+    if (!PtrSCEV)
+      return std::nullopt;
+    SCEVs.push_back(PtrSCEV);
+    if (!PtrSCEVA && !PtrSCEVB) {
+      PtrSCEVA = PtrSCEVB = PtrSCEV;
+      continue;
+    }
+    const SCEV *Diff = SE.getMinusSCEV(PtrSCEV, PtrSCEVA);
+    if (!Diff || isa<SCEVCouldNotCompute>(Diff))
+      return std::nullopt;
+    if (Diff->isNonConstantNegative()) {
+      PtrSCEVA = PtrSCEV;
+      continue;
+    }
+    const SCEV *Diff1 = SE.getMinusSCEV(PtrSCEVB, PtrSCEV);
+    if (!Diff1 || isa<SCEVCouldNotCompute>(Diff1))
+      return std::nullopt;
+    if (Diff1->isNonConstantNegative()) {
+      PtrSCEVB = PtrSCEV;
+      continue;
+    }
+  }
+  const SCEV *Stride = SE.getMinusSCEV(PtrSCEVB, PtrSCEVA);
+  if (!Stride)
+    return std::nullopt;
+  int Size = DL.getTypeStoreSize(ElemTy);
+  auto TryGetStride = [&](const SCEV *Dist,
+                          const SCEV *Multiplier) -> const SCEV * {
+    if (const auto *M = dyn_cast<SCEVMulExpr>(Dist)) {
+      if (M->getOperand(0) == Multiplier)
+        return M->getOperand(1);
+      if (M->getOperand(1) == Multiplier)
+        return M->getOperand(0);
+      return nullptr;
+    }
+    if (Multiplier == Dist)
+      return SE.getConstant(Dist->getType(), 1);
+    return SE.getUDivExactExpr(Dist, Multiplier);
+  };
+  if (Size != 1 || SCEVs.size() > 2) {
+    const SCEV *Sz =
+        SE.getConstant(Stride->getType(), Size * (SCEVs.size() - 1));
+    Stride = TryGetStride(Stride, Sz);
+    if (!Stride)
+      return std::nullopt;
+  }
+  if (!Stride || isa<SCEVConstant>(Stride))
+    return std::nullopt;
+  // Iterate through all pointers and check if all distances are
+  // unique multiple of Dist.
+  using DistOrdPair = std::pair<int64_t, int>;
+  auto Compare = llvm::less_first();
+  std::set<DistOrdPair, decltype(Compare)> Offsets(Compare);
+  int Cnt = 0;
+  bool IsConsecutive = true;
+  for (const SCEV *PtrSCEV : SCEVs) {
+    unsigned Dist = 0;
+    if (PtrSCEV != PtrSCEVA) {
+      const SCEV *Diff = SE.getMinusSCEV(PtrSCEV, PtrSCEVA);
+      const SCEV *Coeff = TryGetStride(Diff, Stride);
+      if (!Coeff)
+        return std::nullopt;
+      const auto *SC = dyn_cast<SCEVConstant>(Coeff);
+      if (!SC || isa<SCEVCouldNotCompute>(SC))
+        return std::nullopt;
+      if (!SE.getMinusSCEV(PtrSCEV,
+                           SE.getAddExpr(PtrSCEVA, SE.getMulExpr(Stride, SC)))
+               ->isZero())
+        return std::nullopt;
+      Dist = SC->getAPInt().getZExtValue();
+    }
+    // If the strides are not the same or repeated, we can't vectorize.
+    if ((Dist / Size) * Size != Dist || (Dist / Size) >= SCEVs.size())
+      return std::nullopt;
+    auto Res = Offsets.emplace(Dist, Cnt);
+    if (!Res.second)
+      return std::nullopt;
+    // Consecutive order if the inserted element is the last one.
+    IsConsecutive = IsConsecutive && std::next(Res.first) == Offsets.end();
+    ++Cnt;
+  }
+  if (Offsets.size() != SCEVs.size())
+    return std::nullopt;
+  SortedIndices.clear();
+  if (!IsConsecutive) {
+    // Fill SortedIndices array only if it is non-consecutive.
+    SortedIndices.resize(PointerOps.size());
+    Cnt = 0;
+    for (const std::pair<int64_t, int> &Pair : Offsets) {
+      SortedIndices[Cnt] = Pair.second;
+      ++Cnt;
+    }
+  }
+  if (!Inst)
+    return nullptr;
+  SCEVExpander Expander(SE, DL, "strided-load-vec");
+  return Expander.expandCodeFor(Stride, Stride->getType(), Inst);
+}
+
 /// Checks if the given array of loads can be represented as a vectorized,
 /// scatter or just simple gather.
 static LoadsState canVectorizeLoads(ArrayRef<Value *> VL, const Value *VL0,
@@ -3927,6 +4044,11 @@ static LoadsState canVectorizeLoads(ArrayRef<Value *> VL, const Value *VL0,
   auto *VecTy = FixedVectorType::get(ScalarTy, Sz);
   // Check the order of pointer operands or that all pointers are the same.
   bool IsSorted = sortPtrAccesses(PointerOps, ScalarTy, DL, SE, Order);
+  Align CommonAlignment = computeCommonAlignment<LoadInst>(VL);
+  if (!IsSorted && Sz > MinProfitableStridedLoads && TTI.isTypeLegal(VecTy) &&
+      TTI.isLegalStridedLoadStore(VecTy, CommonAlignment) &&
+      calculateRtStride(PointerOps, ScalarTy, DL, SE, Order))
+    return LoadsState::StridedVectorize;
   if (IsSorted || all_of(PointerOps, [&](Value *P) {
         return arePointersCompatible(P, PointerOps.front(), TLI);
       })) {
@@ -11645,10 +11767,30 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E, bool PostponedPHIs) {
         std::optional<int> Diff = getPointersDiff(
             VL0->getType(), Ptr0, VL0->getType(), PtrN, *DL, *SE);
         Type *StrideTy = DL->getIndexType(PO->getType());
-        int Stride = *Diff / (static_cast<int>(E->Scalars.size()) - 1);
-        Value *StrideVal =
-            ConstantInt::get(StrideTy, (IsReverseOrder ? -1 : 1) * Stride *
-                                           DL->getTypeAllocSize(ScalarTy));
+        Value *StrideVal;
+        if (Diff) {
+          int Stride = *Diff / (static_cast<int>(E->Scalars.size()) - 1);
+          StrideVal =
+              ConstantInt::get(StrideTy, (IsReverseOrder ? -1 : 1) * Stride *
+                                             DL->getTypeAllocSize(ScalarTy));
+        } else {
+          SmallVector<Value *> PointerOps(E->Scalars.size(), nullptr);
+          transform(E->Scalars, PointerOps.begin(), [](Value *V) {
+            return cast<LoadInst>(V)->getPointerOperand();
+          });
+          OrdersType Order;
+          std::optional<Value *> Stride =
+              calculateRtStride(PointerOps, ScalarTy, *DL, *SE, Order,
+                                &*Builder.GetInsertPoint());
+          Value *NewStride =
+              Builder.CreateIntCast(*Stride, StrideTy, /*isSigned=*/true);
+          StrideVal = Builder.CreateMul(
+              NewStride,
+              ConstantInt::get(
+                  StrideTy,
+                  (IsReverseOrder ? -1 : 1) *
+                      static_cast<int>(DL->getTypeAllocSize(ScalarTy))));
+        }
         Align CommonAlignment = computeCommonAlignment<LoadInst>(E->Scalars);
         auto *Inst = Builder.CreateIntrinsic(
             Intrinsic::experimental_vp_strided_load,
diff --git a/llvm/test/Transforms/SLPVectorizer/RISCV/strided-loads-vectorized.ll b/llvm/test/Transforms/SLPVectorizer/RISCV/strided-loads-vectorized.ll
index 4b0b41970bbb4d..03acc0009fb04c 100644
--- a/llvm/test/Transforms/SLPVectorizer/RISCV/strided-loads-vectorized.ll
+++ b/llvm/test/Transforms/SLPVectorizer/RISCV/strided-loads-vectorized.ll
@@ -78,67 +78,13 @@ define void @test1(ptr %p, ptr noalias %s, i32 %stride) {
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[STR:%.*]] = zext i32 [[STRIDE:%.*]] to i64
 ; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [48 x float], ptr [[P:%.*]], i64 0, i64 0
-; CHECK-NEXT:    [[I:%.*]] = load float, ptr [[ARRAYIDX]], align 4
 ; CHECK-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 30
-; CHECK-NEXT:    [[I1:%.*]] = load float, ptr [[ARRAYIDX1]], align 4
-; CHECK-NEXT:    [[ADD:%.*]] = fsub fast float [[I1]], [[I]]
 ; CHECK-NEXT:    [[ARRAYIDX2:%.*]] = getelementptr inbounds float, ptr [[S:%.*]], i64 0
-; CHECK-NEXT:    store float [[ADD]], ptr [[ARRAYIDX2]], align 4
-; CHECK-NEXT:    [[ARRAYIDX4:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[STR]]
-; CHECK-NEXT:    [[I2:%.*]] = load float, ptr [[ARRAYIDX4]], align 4
-; CHECK-NEXT:    [[ARRAYIDX6:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 26
-; CHECK-NEXT:    [[I3:%.*]] = load float, ptr [[ARRAYIDX6]], align 4
-; CHECK-NEXT:    [[ADD7:%.*]] = fsub fast float [[I3]], [[I2]]
-; CHECK-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds float, ptr [[S]], i64 1
-; CHECK-NEXT:    store float [[ADD7]], ptr [[ARRAYIDX9]], align 4
-; CHECK-NEXT:    [[ST1:%.*]] = mul i64 [[STR]], 2
-; CHECK-NEXT:    [[ARRAYIDX11:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST1]]
-; CHECK-NEXT:    [[I4:%.*]] = load float, ptr [[ARRAYIDX11]], align 4
-; CHECK-NEXT:    [[ARRAYIDX13:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 22
-; CHECK-NEXT:    [[I5:%.*]] = load float, ptr [[ARRAYIDX13]], align 4
-; CHECK-NEXT:    [[ADD14:%.*]] = fsub fast float [[I5]], [[I4]]
-; CHECK-NEXT:    [[ARRAYIDX16:%.*]] = getelementptr inbounds float, ptr [[S]], i64 2
-; CHECK-NEXT:    store float [[ADD14]], ptr [[ARRAYIDX16]], align 4
-; CHECK-NEXT:    [[ST2:%.*]] = mul i64 [[STR]], 3
-; CHECK-NEXT:    [[ARRAYIDX18:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST2]]
-; CHECK-NEXT:    [[I6:%.*]] = load float, ptr [[ARRAYIDX18]], align 4
-; CHECK-NEXT:    [[ARRAYIDX20:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 18
-; CHECK-NEXT:    [[I7:%.*]] = load float, ptr [[ARRAYIDX20]], align 4
-; CHECK-NEXT:    [[ADD21:%.*]] = fsub fast float [[I7]], [[I6]]
-; CHECK-NEXT:    [[ARRAYIDX23:%.*]] = getelementptr inbounds float, ptr [[S]], i64 3
-; CHECK-NEXT:    store float [[ADD21]], ptr [[ARRAYIDX23]], align 4
-; CHECK-NEXT:    [[ST3:%.*]] = mul i64 [[STR]], 4
-; CHECK-NEXT:    [[ARRAYIDX25:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST3]]
-; CHECK-NEXT:    [[I8:%.*]] = load float, ptr [[ARRAYIDX25]], align 4
-; CHECK-NEXT:    [[ARRAYIDX27:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 14
-; CHECK-NEXT:    [[I9:%.*]] = load float, ptr [[ARRAYIDX27]], align 4
-; CHECK-NEXT:    [[ADD28:%.*]] = fsub fast float [[I9]], [[I8]]
-; CHECK-NEXT:    [[ARRAYIDX30:%.*]] = getelementptr inbounds float, ptr [[S]], i64 4
-; CHECK-NEXT:    store float [[ADD28]], ptr [[ARRAYIDX30]], align 4
-; CHECK-NEXT:    [[ST4:%.*]] = mul i64 [[STR]], 5
-; CHECK-NEXT:    [[ARRAYIDX32:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST4]]
-; CHECK-NEXT:    [[I10:%.*]] = load float, ptr [[ARRAYIDX32]], align 4
-; CHECK-NEXT:    [[ARRAYIDX34:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 10
-; CHECK-NEXT:    [[I11:%.*]] = load float, ptr [[ARRAYIDX34]], align 4
-; CHECK-NEXT:    [[ADD35:%.*]] = fsub fast float [[I11]], [[I10]]
-; CHECK-NEXT:    [[ARRAYIDX37:%.*]] = getelementptr inbounds float, ptr [[S]], i64 5
-; CHECK-NEXT:    store float [[ADD35]], ptr [[ARRAYIDX37]], align 4
-; CHECK-NEXT:    [[ST5:%.*]] = mul i64 [[STR]], 6
-; CHECK-NEXT:    [[ARRAYIDX39:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST5]]
-; CHECK-NEXT:    [[I12:%.*]] = load float, ptr [[ARRAYIDX39]], align 4
-; CHECK-NEXT:    [[ARRAYIDX41:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 6
-; CHECK-NEXT:    [[I13:%.*]] = load float, ptr [[ARRAYIDX41]], align 4
-; CHECK-NEXT:    [[ADD42:%.*]] = fsub fast float [[I13]], [[I12]]
-; CHECK-NEXT:    [[ARRAYIDX44:%.*]] = getelementptr inbounds float, ptr [[S]], i64 6
-; CHECK-NEXT:    store float [[ADD42]], ptr [[ARRAYIDX44]], align 4
-; CHECK-NEXT:    [[ST6:%.*]] = mul i64 [[STR]], 7
-; CHECK-NEXT:    [[ARRAYIDX46:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST6]]
-; CHECK-NEXT:    [[I14:%.*]] = load float, ptr [[ARRAYIDX46]], align 4
-; CHECK-NEXT:    [[ARRAYIDX48:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 2
-; CHECK-NEXT:    [[I15:%.*]] = load float, ptr [[ARRAYIDX48]], align 4
-; CHECK-NEXT:    [[ADD49:%.*]] = fsub fast float [[I15]], [[I14]]
-; CHECK-NEXT:    [[ARRAYIDX51:%.*]] = getelementptr inbounds float, ptr [[S]], i64 7
-; CHECK-NEXT:    store float [[ADD49]], ptr [[ARRAYIDX51]], align 4
+; CHECK-NEXT:    [[TMP0:%.*]] = mul i64 [[STR]], 4
+; CHECK-NEXT:    [[TMP1:%.*]] = call <8 x float> @llvm.experimental.vp.strided.load.v8f32.p0.i64(ptr align 4 [[ARRAYIDX]], i64 [[TMP0]], <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, i32 8)
+; CHECK-NEXT:    [[TMP2:%.*]] = call <8 x float> @llvm.experimental.vp.strided.load.v8f32.p0.i64(ptr align 4 [[ARRAYIDX1]], i64 -16, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, i32 8)
+; CHECK-NEXT:    [[TMP3:%.*]] = fsub fast <8 x float> [[TMP2]], [[TMP1]]
+; CHECK-NEXT:    store <8 x float> [[TMP3]], ptr [[ARRAYIDX2]], align 4
 ; CHECK-NEXT:    ret void
 ;
 entry:
@@ -215,38 +161,12 @@ define void @test2(ptr %p, ptr noalias %s, i32 %stride) {
 ; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [48 x float], ptr [[P:%.*]], i64 0, i64 2
 ; CHECK-NEXT:    [[ST6:%.*]] = mul i64 [[STR]], 7
 ; CHECK-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST6]]
-; CHECK-NEXT:    [[I1:%.*]] = load float, ptr [[ARRAYIDX1]], align 4
 ; CHECK-NEXT:    [[ARRAYIDX2:%.*]] = getelementptr inbounds float, ptr [[S:%.*]], i64 0
-; CHECK-NEXT:    [[ST5:%.*]] = mul i64 [[STR]], 6
-; CHECK-NEXT:    [[ARRAYIDX6:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST5]]
-; CHECK-NEXT:    [[I3:%.*]] = load float, ptr [[ARRAYIDX6]], align 4
-; CHECK-NEXT:    [[ST4:%.*]] = mul i64 [[STR]], 5
-; CHECK-NEXT:    [[ARRAYIDX13:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST4]]
-; CHECK-NEXT:    [[I5:%.*]] = load float, ptr [[ARRAYIDX13]], align 4
-; CHECK-NEXT:    [[ST3:%.*]] = mul i64 [[STR]], 4
-; CHECK-NEXT:    [[ARRAYIDX20:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST3]]
-; CHECK-NEXT:    [[I7:%.*]] = load float, ptr [[ARRAYIDX20]], align 4
-; CHECK-NEXT:    [[ST2:%.*]] = mul i64 [[STR]], 3
-; CHECK-NEXT:    [[ARRAYIDX27:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST2]]
-; CHECK-NEXT:    [[I9:%.*]] = load float, ptr [[ARRAYIDX27]], align 4
-; CHECK-NEXT:    [[ST1:%.*]] = mul i64 [[STR]], 2
-; CHECK-NEXT:    [[ARRAYIDX34:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[ST1]]
-; CHECK-NEXT:    [[I11:%.*]] = load float, ptr [[ARRAYIDX34]], align 4
-; CHECK-NEXT:    [[ARRAYIDX41:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 [[STR]]
-; CHECK-NEXT:    [[I13:%.*]] = load float, ptr [[ARRAYIDX41]], align 4
-; CHECK-NEXT:    [[ARRAYIDX48:%.*]] = getelementptr inbounds [48 x float], ptr [[P]], i64 0, i64 0
-; CHECK-NEXT:    [[I15:%.*]] = load float, ptr [[ARRAYIDX48]], align 4
 ; CHECK-NEXT:    [[TMP0:%.*]] = call <8 x float> @llvm.experimental.vp.strided.load.v8f32.p0.i64(ptr align 4 [[ARRAYIDX]], i64 16, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, i32 8)
-; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x float> poison, float [[I1]], i32 0
-; CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x float> [[TMP1]], float [[I3]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x float> [[TMP2]], float [[I5]], i32 2
-; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x float> [[TMP3]], float [[I7]], i32 3
-; CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x float> [[TMP4]], float [[I9]], i32 4
-; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x float> [[TMP5]], float [[I11]], i32 5
-; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x float> [[TMP6]], float [[I13]], i32 6
-; CHECK-NEXT:    [[TMP8:%.*]] = insertelement <8 x float> [[TMP7]], float [[I15]], i32 7
-; CHECK-NEXT:    [[TMP9:%.*]] = fsub fast <8 x float> [[TMP8]], [[TMP0]]
-; CHECK-NEXT:    store <8 x float> [[TMP9]], ptr [[ARRAYIDX2]], align 4
+; CHECK-NEXT:    [[TMP1:%.*]] = mul i64 [[STR]], -4
+; CHECK-NEXT:    [[TMP2:%.*]] = call <8 x float> @llvm.experimental.vp.strided.load.v8f32.p0.i64(ptr align 4 [[ARRAYIDX1]], i64 [[TMP1]], <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, i32 8)
+; CHECK-NEXT:    [[TMP3:%.*]] = fsub fast <8 x float> [[TMP2]], [[TMP0]]
+; CHECK-NEXT:    store <8 x float> [[TMP3]], ptr [[ARRAYIDX2]], align 4
 ; CHECK-NEXT:    ret void
 ;
 entry:



More information about the llvm-commits mailing list