[PATCH] D129763: [LoopVectorize][AArch64] Add TTI hook preferPredicatedReductionSelect

[David Sherwood via Phabricator via llvm-commits]

david-arm created this revision.
david-arm added reviewers: sdesmalen, kmclaughlin, RosieSumpter.
Herald added subscribers: shiva0217, ctetreau, hiraditya, kristof.beyls.
Herald added a project: All.
david-arm requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

By default if SVE is enabled we want the select instruction used for
reductions to be inside the loop, rather than outside. This makes it
possible for the backend to fold the select into the operation to
produce a single predicated add, fadd, etc.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D129763

Files:
  llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
  llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding-reductions.ll


Index: llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding-reductions.ll
===================================================================
--- llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding-reductions.ll
+++ llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding-reductions.ll
@@ -28,14 +28,14 @@
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX1:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT2:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 4 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[VECTOR_PH]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP13:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP14:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX1]], 0
 ; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr i32, i32* [[PTR:%.*]], i64 [[TMP9]]
 ; CHECK-NEXT:    [[TMP11:%.*]] = getelementptr i32, i32* [[TMP10]], i32 0
 ; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i32* [[TMP11]] to <vscale x 4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 4 x i32> @llvm.masked.load.nxv4i32.p0nxv4i32(<vscale x 4 x i32>* [[TMP12]], i32 4, <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i32> poison)
-; CHECK-NEXT:    [[TMP13]] = add <vscale x 4 x i32> [[VEC_PHI]], [[WIDE_MASKED_LOAD]]
-; CHECK-NEXT:    [[TMP14:%.*]] = select <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i32> [[TMP13]], <vscale x 4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[TMP13:%.*]] = add <vscale x 4 x i32> [[VEC_PHI]], [[WIDE_MASKED_LOAD]]
+; CHECK-NEXT:    [[TMP14]] = select <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i32> [[TMP13]], <vscale x 4 x i32> [[VEC_PHI]]
 ; CHECK-NEXT:    [[TMP15:%.*]] = call i64 @llvm.vscale.i64()
 ; CHECK-NEXT:    [[TMP16:%.*]] = mul i64 [[TMP15]], 4
 ; CHECK-NEXT:    [[INDEX_NEXT2]] = add i64 [[INDEX1]], [[TMP16]]
@@ -225,7 +225,7 @@
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 4 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[VECTOR_PH]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ insertelement (<vscale x 4 x i32> zeroinitializer, i32 7, i32 0), [[VECTOR_PH]] ], [ [[PREDPHI:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ insertelement (<vscale x 4 x i32> zeroinitializer, i32 7, i32 0), [[VECTOR_PH]] ], [ [[TMP21:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 0
 ; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[COND:%.*]], i64 [[TMP9]]
 ; CHECK-NEXT:    [[TMP11:%.*]] = getelementptr inbounds i32, i32* [[TMP10]], i32 0
@@ -240,8 +240,8 @@
 ; CHECK-NEXT:    [[TMP18:%.*]] = xor <vscale x 4 x i32> [[VEC_PHI]], [[WIDE_MASKED_LOAD1]]
 ; CHECK-NEXT:    [[TMP19:%.*]] = xor <vscale x 4 x i1> [[TMP13]], shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i32 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer)
 ; CHECK-NEXT:    [[TMP20:%.*]] = select <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i1> [[TMP19]], <vscale x 4 x i1> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI]] = select <vscale x 4 x i1> [[TMP15]], <vscale x 4 x i32> [[TMP18]], <vscale x 4 x i32> [[VEC_PHI]]
-; CHECK-NEXT:    [[TMP21:%.*]] = select <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i32> [[PREDPHI]], <vscale x 4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <vscale x 4 x i1> [[TMP15]], <vscale x 4 x i32> [[TMP18]], <vscale x 4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[TMP21]] = select <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i32> [[PREDPHI]], <vscale x 4 x i32> [[VEC_PHI]]
 ; CHECK-NEXT:    [[TMP22:%.*]] = call i64 @llvm.vscale.i64()
 ; CHECK-NEXT:    [[TMP23:%.*]] = mul i64 [[TMP22]], 4
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP23]]
Index: llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
===================================================================
--- llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
+++ llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
@@ -347,6 +347,11 @@
   bool isLegalToVectorizeReduction(const RecurrenceDescriptor &RdxDesc,
                                    ElementCount VF) const;
 
+  bool preferPredicatedReductionSelect(unsigned Opcode, Type *Ty,
+                                       TTI::ReductionFlags Flags) const {
+    return ST->hasSVE();
+  }
+
   InstructionCost getArithmeticReductionCost(unsigned Opcode, VectorType *Ty,
                                              Optional<FastMathFlags> FMF,
                                              TTI::TargetCostKind CostKind);


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D129763.444612.patch
Type: text/x-patch
Size: 4947 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220714/2f5c9ffc/attachment.bin>