[LoopVectorizer] Propagate underlying instruction to the cloned instances of VPPartialReductionRecipes

llvm/lib/Transforms/Vectorize/VPlan.h

@@ -2459,7 +2459,8 @@ class VPPartialReductionRecipe : public VPSingleDefRecipe {
   ~VPPartialReductionRecipe() override = default;
 
   VPPartialReductionRecipe *clone() override {
-    return new VPPartialReductionRecipe(Opcode, getOperand(0), getOperand(1));
+    return new VPPartialReductionRecipe(Opcode, getOperand(0), getOperand(1),
+                                        getUnderlyingInstr());
   }
 
   VP_CLASSOF_IMPL(VPDef::VPPartialReductionSC)

llvm/test/Transforms/LoopVectorize/AArch64/vplan-printing.ll

@@ -1,5 +1,5 @@
 ; REQUIRES: asserts
-; RUN: opt -mattr=+neon,+dotprod -passes=loop-vectorize -debug-only=loop-vectorize -force-vector-interleave=1 -disable-output %s 2>&1 | FileCheck %s
+; RUN: opt -mattr=+neon,+dotprod -passes=loop-vectorize -debug-only=loop-vectorize -force-vector-interleave=1 -enable-epilogue-vectorization -epilogue-vectorization-force-VF=2 -disable-output %s 2>&1 | FileCheck %s
 
 target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128"
 target triple = "aarch64-none-unknown-elf"
@@ -70,7 +70,71 @@ define i32 @print_partial_reduction(ptr %a, ptr %b) {
 ; CHECK-NEXT:   IR   %add.lcssa = phi i32 [ %add, %for.body ] (extra operand: vp<[[EXTRACT]]> from middle.block)
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
-;
+; CHECK: VPlan 'Final VPlan for VF={8,16},UF={1}' {
+; CHECK-NEXT: Live-in ir<[[EP_VFxUF:.+]]> = VF * UF
+; CHECK-NEXT: Live-in ir<[[EP_VEC_TC:.+]]> = vector-trip-count
+; CHECK-NEXT: Live-in ir<1024> = original trip-count
+; CHECK-EMPTY:
+; CHECK-NEXT: ir-bb<entry>:
+; CHECK-NEXT: Successor(s): ir-bb<scalar.ph>, ir-bb<vector.main.loop.iter.check>
+; CHECK-EMPTY:
+; CHECK-NEXT: ir-bb<vector.main.loop.iter.check>:
+; CHECK-NEXT: Successor(s): ir-bb<scalar.ph>, ir-bb<vector.ph>
+; CHECK-EMPTY:
+; CHECK-NEXT: ir-bb<vector.ph>:
+; CHECK-NEXT: Successor(s): vector loop
+; CHECK-EMPTY:
+; CHECK-NEXT: <x1> vector loop: {
+; CHECK-NEXT:   vector.body:
+; CHECK-NEXT:     SCALAR-PHI vp<[[EP_IV:%.+]]> = phi ir<0>, vp<%index.next>
+; CHECK-NEXT:     WIDEN-REDUCTION-PHI ir<%accum> = phi ir<0>, ir<%add> (VF scaled by 1/4)
+; CHECK-NEXT:     vp<[[STEPS:%.+]]> = SCALAR-STEPS vp<[[EP_IV]]>, ir<1>
+; CHECK-NEXT:     CLONE ir<%gep.a> = getelementptr ir<%a>, vp<[[STEPS]]>
+; CHECK-NEXT:     vp<[[PTR_A:%.+]]> = vector-pointer ir<%gep.a>
+; CHECK-NEXT:     WIDEN ir<%load.a> = load vp<[[PTR_A]]>
+; CHECK-NEXT:     WIDEN-CAST ir<%ext.a> = zext ir<%load.a> to i32
+; CHECK-NEXT:     CLONE ir<%gep.b> = getelementptr ir<%b>, vp<[[STEPS]]>
+; CHECK-NEXT:     vp<[[PTR_B:%.+]]> = vector-pointer ir<%gep.b>
+; CHECK-NEXT:     WIDEN ir<%load.b> = load vp<[[PTR_B]]>
+; CHECK-NEXT:     WIDEN-CAST ir<%ext.b> = zext ir<%load.b> to i32
+; CHECK-NEXT:     WIDEN ir<%mul> = mul ir<%ext.b>, ir<%ext.a>
+; CHECK-NEXT:     PARTIAL-REDUCE ir<%add> = add ir<%mul>, ir<%accum>
+; CHECK-NEXT:     EMIT vp<[[EP_IV_NEXT:%.+]]> = add nuw vp<[[EP_IV]]>, ir<16>
+; CHECK-NEXT:     EMIT branch-on-count vp<[[EP_IV_NEXT]]>, ir<1024>
+; CHECK-NEXT:   No successors
+; CHECK-NEXT: }
+; CHECK-NEXT: Successor(s): ir-bb<middle.block>
+; CHECK-EMPTY:
+; CHECK-NEXT: ir-bb<middle.block>:
+; CHECK-NEXT:   EMIT vp<[[RED_RESULT:%.+]]> = compute-reduction-result ir<%accum>, ir<%add>
+; CHECK-NEXT:   EMIT vp<[[EXTRACT:%.+]]> = extract-from-end vp<[[RED_RESULT]]>, ir<1>
+; CHECK-NEXT:   EMIT vp<[[CMP:%.+]]> = icmp eq ir<1024>, ir<1024>
+; CHECK-NEXT:   EMIT branch-on-cond vp<[[CMP]]>
+; CHECK-NEXT: Successor(s): ir-bb<exit>, ir-bb<scalar.ph>
+; CHECK-EMPTY:
+; CHECK-NEXT: ir-bb<exit>:
+; CHECK-NEXT:   IR   %add.lcssa = phi i32 [ %add, %for.body ] (extra operand: vp<[[EXTRACT]]> from ir-bb<middle.block>)
+; CHECK-NEXT: No successors
+; CHECK-EMPTY:
+; CHECK-NEXT: ir-bb<scalar.ph>:
+; CHECK-NEXT:   EMIT vp<[[EP_RESUME:%.+]]> = resume-phi ir<1024>, ir<0>
+; CHECK-NEXT:   EMIT vp<[[EP_MERGE:%.+]]> = resume-phi vp<[[RED_RESULT]]>, ir<0>
+; CHECK-NEXT: Successor(s): ir-bb<for.body>
+; CHECK-EMPTY:
+; CHECK-NEXT: ir-bb<for.body>:
+; CHECK-NEXT:   IR   %accum = phi i32 [ 0, %scalar.ph ], [ %add, %for.body ] (extra operand: vp<[[EP_MERGE]]> from ir-bb<scalar.ph>)
+; CHECK-NEXT:   IR   %gep.a = getelementptr i8, ptr %a, i64 %iv
+; CHECK-NEXT:   IR   %load.a = load i8, ptr %gep.a, align 1
+; CHECK-NEXT:   IR   %ext.a = zext i8 %load.a to i32
+; CHECK-NEXT:   IR   %gep.b = getelementptr i8, ptr %b, i64 %iv
+; CHECK-NEXT:   IR   %load.b = load i8, ptr %gep.b, align 1
+; CHECK-NEXT:   IR   %ext.b = zext i8 %load.b to i32
+; CHECK-NEXT:   IR   %mul = mul i32 %ext.b, %ext.a
+; CHECK-NEXT:   IR   %add = add i32 %mul, %accum
+; CHECK-NEXT:   IR   %iv.next = add i64 %iv, 1
+; CHECK-NEXT:   IR   %exitcond.not = icmp eq i64 %iv.next, 1024
+; CHECK-NEXT: No successors
+; CHECK-NEXT: }
 entry:
   br label %for.body