[llvm] 1189635 - [VPlan] Add VPInterleaveRecipe::NeedsMaskForGaps field (NFCI).

[Florian Hahn via llvm-commits]

Author: Florian Hahn
Date: 2023-04-07T13:11:03+01:00
New Revision: 11896357d413553a705ea552dbeeaa591c8f66ca

URL: https://github.com/llvm/llvm-project/commit/11896357d413553a705ea552dbeeaa591c8f66ca
DIFF: https://github.com/llvm/llvm-project/commit/11896357d413553a705ea552dbeeaa591c8f66ca.diff

LOG: [VPlan] Add VPInterleaveRecipe::NeedsMaskForGaps field (NFCI).

This patch adds a NeedsMaskForGaps field to VPInterleaveRecipe to record
whether a mask for gaps is needed. This removes a dependence on the cost
model in VPlan code-generation.

Reviewed By: Ayal

Differential Revision: https://reviews.llvm.org/D147467

Added: 
    

Modified: 
    llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
    llvm/lib/Transforms/Vectorize/VPlan.h
    llvm/unittests/Transforms/Vectorize/VPlanTest.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
index 51ea1a70caa57..9c822b4a5301d 100644
--- a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -535,7 +535,7 @@ class InnerLoopVectorizer {
                                 ArrayRef<VPValue *> VPDefs,
                                 VPTransformState &State, VPValue *Addr,
                                 ArrayRef<VPValue *> StoredValues,
-                                VPValue *BlockInMask = nullptr);
+                                VPValue *BlockInMask, bool NeedsMaskForGaps);
 
   /// Fix the non-induction PHIs in \p Plan.
   void fixNonInductionPHIs(VPlan &Plan, VPTransformState &State);
@@ -2610,7 +2610,7 @@ static bool useMaskedInterleavedAccesses(const TargetTransformInfo &TTI) {
 void InnerLoopVectorizer::vectorizeInterleaveGroup(
     const InterleaveGroup<Instruction> *Group, ArrayRef<VPValue *> VPDefs,
     VPTransformState &State, VPValue *Addr, ArrayRef<VPValue *> StoredValues,
-    VPValue *BlockInMask) {
+    VPValue *BlockInMask, bool NeedsMaskForGaps) {
   Instruction *Instr = Group->getInsertPos();
   const DataLayout &DL = Instr->getModule()->getDataLayout();
 
@@ -2671,7 +2671,7 @@ void InnerLoopVectorizer::vectorizeInterleaveGroup(
   // Vectorize the interleaved load group.
   if (isa<LoadInst>(Instr)) {
     Value *MaskForGaps = nullptr;
-    if (Group->requiresScalarEpilogue() && !Cost->isScalarEpilogueAllowed()) {
+    if (NeedsMaskForGaps) {
       MaskForGaps =
           createBitMaskForGaps(Builder, VF.getKnownMinValue(), *Group);
       assert(MaskForGaps && "Mask for Gaps is required but it is null");
@@ -9033,8 +9033,10 @@ VPlanPtr LoopVectorizationPlanner::buildVPlanWithVPRecipes(
         StoredValues.push_back(StoreR->getStoredValue());
       }
 
+    bool NeedsMaskForGaps =
+        IG->requiresScalarEpilogue() && !CM.isScalarEpilogueAllowed();
     auto *VPIG = new VPInterleaveRecipe(IG, Recipe->getAddr(), StoredValues,
-                                        Recipe->getMask());
+                                        Recipe->getMask(), NeedsMaskForGaps);
     VPIG->insertBefore(Recipe);
     unsigned J = 0;
     for (unsigned i = 0; i < IG->getFactor(); ++i)
@@ -9487,7 +9489,8 @@ void VPScalarIVStepsRecipe::execute(VPTransformState &State) {
 void VPInterleaveRecipe::execute(VPTransformState &State) {
   assert(!State.Instance && "Interleave group being replicated.");
   State.ILV->vectorizeInterleaveGroup(IG, definedValues(), State, getAddr(),
-                                      getStoredValues(), getMask());
+                                      getStoredValues(), getMask(),
+                                      NeedsMaskForGaps);
 }
 
 void VPReductionRecipe::execute(VPTransformState &State) {

diff  --git a/llvm/lib/Transforms/Vectorize/VPlan.h b/llvm/lib/Transforms/Vectorize/VPlan.h
index 77a32d854be14..596207bd0668c 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.h
+++ b/llvm/lib/Transforms/Vectorize/VPlan.h
@@ -1391,12 +1391,20 @@ class VPBlendRecipe : public VPRecipeBase, public VPValue {
 class VPInterleaveRecipe : public VPRecipeBase {
   const InterleaveGroup<Instruction> *IG;
 
+  /// Indicates if the interleave group is in a conditional block and requires a
+  /// mask.
   bool HasMask = false;
 
+  /// Indicates if gaps between members of the group need to be masked out or if
+  /// unusued gaps can be loaded speculatively.
+  bool NeedsMaskForGaps = false;
+
 public:
   VPInterleaveRecipe(const InterleaveGroup<Instruction> *IG, VPValue *Addr,
-                     ArrayRef<VPValue *> StoredValues, VPValue *Mask)
-      : VPRecipeBase(VPDef::VPInterleaveSC, {Addr}), IG(IG) {
+                     ArrayRef<VPValue *> StoredValues, VPValue *Mask,
+                     bool NeedsMaskForGaps)
+      : VPRecipeBase(VPDef::VPInterleaveSC, {Addr}), IG(IG),
+        NeedsMaskForGaps(NeedsMaskForGaps) {
     for (unsigned i = 0; i < IG->getFactor(); ++i)
       if (Instruction *I = IG->getMember(i)) {
         if (I->getType()->isVoidTy())

diff  --git a/llvm/unittests/Transforms/Vectorize/VPlanTest.cpp b/llvm/unittests/Transforms/Vectorize/VPlanTest.cpp
index 9a150cffae502..903e8232a617d 100644
--- a/llvm/unittests/Transforms/Vectorize/VPlanTest.cpp
+++ b/llvm/unittests/Transforms/Vectorize/VPlanTest.cpp
@@ -960,7 +960,7 @@ TEST(VPRecipeTest, CastVPInterleaveRecipeToVPUser) {
   VPValue Addr;
   VPValue Mask;
   InterleaveGroup<Instruction> IG(4, false, Align(4));
-  VPInterleaveRecipe Recipe(&IG, &Addr, {}, &Mask);
+  VPInterleaveRecipe Recipe(&IG, &Addr, {}, &Mask, false);
   EXPECT_TRUE(isa<VPUser>(&Recipe));
   VPRecipeBase *BaseR = &Recipe;
   EXPECT_TRUE(isa<VPUser>(BaseR));