[llvm] [LV] Generate scalar VPInstruction::Not if possible. (PR #152628)
via llvm-commits
llvm-commits at lists.llvm.org
Thu Aug 7 19:50:07 PDT 2025
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-llvm-transforms
Author: Elvis Wang (ElvisWang123)
<details>
<summary>Changes</summary>
`VPInstruction::Not` which will generate xor instruction is widely used for the exit condition. This patch make `VPInstruction::Not` generate scalar `xor` if possible.
This can help reducing the (splat true) in the `xor` and make `xor` be scalar.
---
Patch is 146.72 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/152628.diff
28 Files Affected:
- (modified) llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp (+3-1)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/conditional-branches-cost.ll (+5-5)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/divs-with-scalable-vfs.ll (+6-6)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/first-order-recurrence-fold-tail.ll (+3-3)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/fmax-without-fast-math-flags.ll (+1-2)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/fmin-without-fast-math-flags.ll (+1-2)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/induction-costs-sve.ll (+9-9)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/low_trip_count_predicates.ll (+3-3)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll (+5-5)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/masked-call.ll (+24-24)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/optsize_minsize.ll (+9-9)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/partial-reduce-dot-product.ll (+6-6)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/partial-reduce.ll (+6-6)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/reduction-recurrence-costs-sve.ll (+6-6)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/scalable-strict-fadd.ll (+14-14)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/store-costs-sve.ll (+5-5)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/sve-interleaved-masked-accesses.ll (+185-185)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/sve-low-trip-count.ll (+2-2)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding-forced.ll (+3-3)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding-optsize.ll (+3-3)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding-reductions.ll (+13-13)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding-unroll.ll (+6-6)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding.ll (+28-28)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/tail-fold-uniform-memops.ll (+2-2)
- (modified) llvm/test/Transforms/LoopVectorize/AArch64/tail-folding-styles.ll (+6-6)
- (modified) llvm/test/Transforms/LoopVectorize/fmax-without-fast-math-flags-interleave.ll (+1-2)
- (modified) llvm/test/Transforms/LoopVectorize/fmax-without-fast-math-flags.ll (+4-8)
- (modified) llvm/test/Transforms/LoopVectorize/fmin-without-fast-math-flags.ll (+2-4)
``````````diff
diff --git a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
index e971ba1aac15c..6f476a41aa63d 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
@@ -517,6 +517,7 @@ bool VPInstruction::canGenerateScalarForFirstLane() const {
case VPInstruction::PtrAdd:
case VPInstruction::ExplicitVectorLength:
case VPInstruction::AnyOf:
+ case VPInstruction::Not:
return true;
default:
return false;
@@ -569,7 +570,8 @@ Value *VPInstruction::generate(VPTransformState &State) {
switch (getOpcode()) {
case VPInstruction::Not: {
- Value *A = State.get(getOperand(0));
+ bool OnlyFirstLaneUsed = vputils::onlyFirstLaneUsed(this);
+ Value *A = State.get(getOperand(0), OnlyFirstLaneUsed);
return Builder.CreateNot(A, Name);
}
case Instruction::ExtractElement: {
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/conditional-branches-cost.ll b/llvm/test/Transforms/LoopVectorize/AArch64/conditional-branches-cost.ll
index 0232d88347d0a..1187bcf9ca658 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/conditional-branches-cost.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/conditional-branches-cost.ll
@@ -728,9 +728,9 @@ define void @multiple_exit_conditions(ptr %src, ptr noalias %dst) #1 {
; PRED-NEXT: call void @llvm.masked.store.nxv2f64.p0(<vscale x 2 x double> [[TMP14]], ptr [[NEXT_GEP]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; PRED-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP5]]
; PRED-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX]], i64 [[TMP10]])
-; PRED-NEXT: [[TMP16:%.*]] = xor <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
-; PRED-NEXT: [[TMP17:%.*]] = extractelement <vscale x 2 x i1> [[TMP16]], i32 0
-; PRED-NEXT: br i1 [[TMP17]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
+; PRED-NEXT: [[TMP15:%.*]] = extractelement <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; PRED-NEXT: [[TMP16:%.*]] = xor i1 [[TMP15]], true
+; PRED-NEXT: br i1 [[TMP16]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; PRED: [[MIDDLE_BLOCK]]:
; PRED-NEXT: br label %[[EXIT:.*]]
; PRED: [[SCALAR_PH]]:
@@ -1394,9 +1394,9 @@ define void @test_conditional_interleave_group (ptr noalias %src.1, ptr noalias
; PRED: [[PRED_STORE_CONTINUE27]]:
; PRED-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 8
; PRED-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <8 x i1> @llvm.get.active.lane.mask.v8i1.i64(i64 [[INDEX]], i64 [[TMP17]])
-; PRED-NEXT: [[TMP84:%.*]] = xor <8 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; PRED-NEXT: [[TMP84:%.*]] = extractelement <8 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; PRED-NEXT: [[TMP85:%.*]] = xor i1 [[TMP84]], true
; PRED-NEXT: [[VEC_IND_NEXT]] = add <8 x i64> [[VEC_IND]], splat (i64 8)
-; PRED-NEXT: [[TMP85:%.*]] = extractelement <8 x i1> [[TMP84]], i32 0
; PRED-NEXT: br i1 [[TMP85]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
; PRED: [[MIDDLE_BLOCK]]:
; PRED-NEXT: br label %[[EXIT:.*]]
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/divs-with-scalable-vfs.ll b/llvm/test/Transforms/LoopVectorize/AArch64/divs-with-scalable-vfs.ll
index 1ad1e42678c5a..bae979d46c7f3 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/divs-with-scalable-vfs.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/divs-with-scalable-vfs.ll
@@ -151,10 +151,10 @@ define void @sdiv_feeding_gep_predicated(ptr %dst, i32 %x, i64 %M, i64 %conv6, i
; CHECK-NEXT: call void @llvm.masked.store.nxv2f64.p0(<vscale x 2 x double> zeroinitializer, ptr [[TMP34]], i32 8, <vscale x 2 x i1> [[TMP23]])
; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
; CHECK-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX]], i64 [[TMP14]])
-; CHECK-NEXT: [[TMP36:%.*]] = xor <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; CHECK-NEXT: [[TMP35:%.*]] = extractelement <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; CHECK-NEXT: [[TMP36:%.*]] = xor i1 [[TMP35]], true
; CHECK-NEXT: [[VEC_IND_NEXT]] = add <vscale x 2 x i64> [[VEC_IND]], [[DOTSPLAT]]
-; CHECK-NEXT: [[TMP37:%.*]] = extractelement <vscale x 2 x i1> [[TMP36]], i32 0
-; CHECK-NEXT: br i1 [[TMP37]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
+; CHECK-NEXT: br i1 [[TMP36]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
; CHECK-NEXT: br label %[[EXIT:.*]]
; CHECK: [[SCALAR_PH]]:
@@ -276,10 +276,10 @@ define void @udiv_urem_feeding_gep(i64 %x, ptr %dst, i64 %N) {
; CHECK-NEXT: call void @llvm.masked.store.nxv2i64.p0(<vscale x 2 x i64> [[TMP23]], ptr [[TMP38]], i32 4, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
; CHECK-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX]], i64 [[TMP14]])
-; CHECK-NEXT: [[TMP47:%.*]] = xor <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; CHECK-NEXT: [[TMP39:%.*]] = extractelement <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; CHECK-NEXT: [[TMP40:%.*]] = xor i1 [[TMP39]], true
; CHECK-NEXT: [[VEC_IND_NEXT]] = add <vscale x 2 x i64> [[VEC_IND]], [[BROADCAST_SPLAT4]]
-; CHECK-NEXT: [[TMP48:%.*]] = extractelement <vscale x 2 x i1> [[TMP47]], i32 0
-; CHECK-NEXT: br i1 [[TMP48]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
+; CHECK-NEXT: br i1 [[TMP40]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
; CHECK-NEXT: br label %[[EXIT:.*]]
; CHECK: [[SCALAR_PH]]:
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/first-order-recurrence-fold-tail.ll b/llvm/test/Transforms/LoopVectorize/AArch64/first-order-recurrence-fold-tail.ll
index fff99f1498ae7..75857ab9c030d 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/first-order-recurrence-fold-tail.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/first-order-recurrence-fold-tail.ll
@@ -64,10 +64,10 @@ define i32 @test_phi_iterator_invalidation(ptr %A, ptr noalias %B) {
; CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[TMP26]], ptr [[TMP28]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 4
; CHECK-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i64(i64 [[INDEX_NEXT]], i64 1002)
-; CHECK-NEXT: [[TMP30:%.*]] = xor <4 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; CHECK-NEXT: [[TMP29:%.*]] = extractelement <4 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; CHECK-NEXT: [[TMP30:%.*]] = xor i1 [[TMP29]], true
; CHECK-NEXT: [[VEC_IND_NEXT]] = add <4 x i64> [[VEC_IND]], splat (i64 4)
-; CHECK-NEXT: [[TMP31:%.*]] = extractelement <4 x i1> [[TMP30]], i32 0
-; CHECK-NEXT: br i1 [[TMP31]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT: br i1 [[TMP30]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: br label [[EXIT:%.*]]
; CHECK: scalar.ph:
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/fmax-without-fast-math-flags.ll b/llvm/test/Transforms/LoopVectorize/AArch64/fmax-without-fast-math-flags.ll
index 5de9d0e3fc93f..3475b951e54f5 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/fmax-without-fast-math-flags.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/fmax-without-fast-math-flags.ll
@@ -75,8 +75,7 @@ define float @fmaxnum(ptr %src, i64 %n) {
; CHECK-NEXT: [[RDX_MINMAX_SELECT:%.*]] = call <4 x float> @llvm.maxnum.v4f32(<4 x float> [[TMP11]], <4 x float> [[TMP12]])
; CHECK-NEXT: [[TMP13:%.*]] = call float @llvm.vector.reduce.fmax.v4f32(<4 x float> [[RDX_MINMAX_SELECT]])
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
-; CHECK-NEXT: [[TMP15:%.*]] = xor <4 x i1> [[BROADCAST_SPLAT]], splat (i1 true)
-; CHECK-NEXT: [[TMP16:%.*]] = extractelement <4 x i1> [[TMP15]], i32 0
+; CHECK-NEXT: [[TMP16:%.*]] = xor i1 [[TMP6]], true
; CHECK-NEXT: [[TMP17:%.*]] = and i1 [[CMP_N]], [[TMP16]]
; CHECK-NEXT: br i1 [[TMP17]], label %[[EXIT:.*]], label %[[SCALAR_PH]]
; CHECK: [[SCALAR_PH]]:
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/fmin-without-fast-math-flags.ll b/llvm/test/Transforms/LoopVectorize/AArch64/fmin-without-fast-math-flags.ll
index ea44fc35e1484..ab4ec848ad4c5 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/fmin-without-fast-math-flags.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/fmin-without-fast-math-flags.ll
@@ -75,8 +75,7 @@ define float @fminnum(ptr %src, i64 %n) {
; CHECK-NEXT: [[RDX_MINMAX_SELECT:%.*]] = call <4 x float> @llvm.minnum.v4f32(<4 x float> [[TMP11]], <4 x float> [[TMP12]])
; CHECK-NEXT: [[TMP13:%.*]] = call float @llvm.vector.reduce.fmin.v4f32(<4 x float> [[RDX_MINMAX_SELECT]])
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
-; CHECK-NEXT: [[TMP15:%.*]] = xor <4 x i1> [[BROADCAST_SPLAT]], splat (i1 true)
-; CHECK-NEXT: [[TMP16:%.*]] = extractelement <4 x i1> [[TMP15]], i32 0
+; CHECK-NEXT: [[TMP16:%.*]] = xor i1 [[TMP6]], true
; CHECK-NEXT: [[TMP17:%.*]] = and i1 [[CMP_N]], [[TMP16]]
; CHECK-NEXT: br i1 [[TMP17]], label %[[EXIT:.*]], label %[[SCALAR_PH]]
; CHECK: [[SCALAR_PH]]:
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/induction-costs-sve.ll b/llvm/test/Transforms/LoopVectorize/AArch64/induction-costs-sve.ll
index 8b354d91909b1..cbe1aaac0cb81 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/induction-costs-sve.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/induction-costs-sve.ll
@@ -134,9 +134,9 @@ define void @iv_casts(ptr %dst, ptr %src, i32 %x, i64 %N) #0 {
; PRED-NEXT: call void @llvm.masked.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP23]], ptr [[TMP26]], i32 1, <vscale x 16 x i1> [[ACTIVE_LANE_MASK]])
; PRED-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
; PRED-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 16 x i1> @llvm.get.active.lane.mask.nxv16i1.i64(i64 [[INDEX]], i64 [[TMP15]])
-; PRED-NEXT: [[TMP28:%.*]] = xor <vscale x 16 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
-; PRED-NEXT: [[TMP29:%.*]] = extractelement <vscale x 16 x i1> [[TMP28]], i32 0
-; PRED-NEXT: br i1 [[TMP29]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; PRED-NEXT: [[TMP25:%.*]] = extractelement <vscale x 16 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; PRED-NEXT: [[TMP27:%.*]] = xor i1 [[TMP25]], true
+; PRED-NEXT: br i1 [[TMP27]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; PRED: [[MIDDLE_BLOCK]]:
; PRED-NEXT: br label %[[EXIT:.*]]
; PRED: [[SCALAR_PH]]:
@@ -304,9 +304,9 @@ define void @iv_trunc(i32 %x, ptr %dst, i64 %N) #0 {
; PRED: [[PRED_STORE_CONTINUE2]]:
; PRED-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 2
; PRED-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 [[INDEX]], i64 [[TMP15]])
-; PRED-NEXT: [[TMP24:%.*]] = xor <2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; PRED-NEXT: [[TMP24:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; PRED-NEXT: [[TMP25:%.*]] = xor i1 [[TMP24]], true
; PRED-NEXT: [[VEC_IND_NEXT]] = add <2 x i32> [[VEC_IND]], splat (i32 2)
-; PRED-NEXT: [[TMP25:%.*]] = extractelement <2 x i1> [[TMP24]], i32 0
; PRED-NEXT: br i1 [[TMP25]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; PRED: [[MIDDLE_BLOCK]]:
; PRED-NEXT: br label %[[EXIT:.*]]
@@ -496,9 +496,9 @@ define void @trunc_ivs_and_store(i32 %x, ptr %dst, i64 %N) #0 {
; PRED: [[PRED_STORE_CONTINUE7]]:
; PRED-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 4
; PRED-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i64(i64 [[INDEX]], i64 [[TMP16]])
-; PRED-NEXT: [[TMP35:%.*]] = xor <4 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; PRED-NEXT: [[TMP35:%.*]] = extractelement <4 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; PRED-NEXT: [[TMP36:%.*]] = xor i1 [[TMP35]], true
; PRED-NEXT: [[VEC_IND_NEXT]] = add <4 x i32> [[VEC_IND]], splat (i32 4)
-; PRED-NEXT: [[TMP36:%.*]] = extractelement <4 x i1> [[TMP35]], i32 0
; PRED-NEXT: br i1 [[TMP36]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; PRED: [[MIDDLE_BLOCK]]:
; PRED-NEXT: br label %[[EXIT:.*]]
@@ -690,9 +690,9 @@ define void @ivs_trunc_and_ext(i32 %x, ptr %dst, i64 %N) #0 {
; PRED: [[PRED_STORE_CONTINUE6]]:
; PRED-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 4
; PRED-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i64(i64 [[INDEX]], i64 [[TMP15]])
-; PRED-NEXT: [[TMP34:%.*]] = xor <4 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; PRED-NEXT: [[TMP34:%.*]] = extractelement <4 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; PRED-NEXT: [[TMP35:%.*]] = xor i1 [[TMP34]], true
; PRED-NEXT: [[VEC_IND_NEXT]] = add <4 x i32> [[VEC_IND]], splat (i32 4)
-; PRED-NEXT: [[TMP35:%.*]] = extractelement <4 x i1> [[TMP34]], i32 0
; PRED-NEXT: br i1 [[TMP35]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
; PRED: [[MIDDLE_BLOCK]]:
; PRED-NEXT: br label %[[EXIT:.*]]
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/low_trip_count_predicates.ll b/llvm/test/Transforms/LoopVectorize/AArch64/low_trip_count_predicates.ll
index 5f5d326cb4bba..479a9c979f3e6 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/low_trip_count_predicates.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/low_trip_count_predicates.ll
@@ -436,9 +436,9 @@ define void @overflow_indvar_known_false(ptr nocapture noundef %p, i32 noundef %
; CHECK-NEXT: call void @llvm.masked.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr [[TMP13]], i32 1, <vscale x 16 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP8]]
; CHECK-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 16 x i1> @llvm.get.active.lane.mask.nxv16i1.i64(i64 [[INDEX_NEXT]], i64 [[TMP1]])
-; CHECK-NEXT: [[TMP16:%.*]] = xor <vscale x 16 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
-; CHECK-NEXT: [[TMP17:%.*]] = extractelement <vscale x 16 x i1> [[TMP16]], i32 0
-; CHECK-NEXT: br i1 [[TMP17]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
+; CHECK-NEXT: [[TMP30:%.*]] = extractelement <vscale x 16 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; CHECK-NEXT: [[TMP31:%.*]] = xor i1 [[TMP30]], true
+; CHECK-NEXT: br i1 [[TMP31]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
; CHECK-NEXT: br label %[[WHILE_END_LOOPEXIT:.*]]
; CHECK: [[SCALAR_PH]]:
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll b/llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll
index 078f98f54525b..aed1c3d9fcc4c 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll
@@ -87,8 +87,8 @@ define void @test_widen_exp_v2(ptr noalias %p2, ptr noalias %p, i64 %n) #5 {
; TFCOMMON: pred.store.continue2:
; TFCOMMON-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 2
; TFCOMMON-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 [[INDEX]], i64 [[TMP3]])
-; TFCOMMON-NEXT: [[TMP18:%.*]] = xor <2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
-; TFCOMMON-NEXT: [[TMP17:%.*]] = extractelement <2 x i1> [[TMP18]], i32 0
+; TFCOMMON-NEXT: [[TMP15:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; TFCOMMON-NEXT: [[TMP17:%.*]] = xor i1 [[TMP15]], true
; TFCOMMON-NEXT: br i1 [[TMP17]], label [[END:%.*]], label [[LOOP]], !llvm.loop [[LOOP0:![0-9]+]]
; TFCOMMON: end:
; TFCOMMON-NEXT: ret void
@@ -153,9 +153,9 @@ define void @test_widen_exp_v2(ptr noalias %p2, ptr noalias %p, i64 %n) #5 {
; TFA_INTERLEAVE-NEXT: [[TMP27:%.*]] = add i64 [[INDEX]], 2
; TFA_INTERLEAVE-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 [[INDEX]], i64 [[TMP3]])
; TFA_INTERLEAVE-NEXT: [[ACTIVE_LANE_MASK_NEXT10]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 [[TMP27]], i64 [[TMP3]])
-; TFA_INTERLEAVE-NEXT: [[TMP28:%.*]] = xor <2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
-; TFA_INTERLEAVE-NEXT: [[TMP30:%.*]] = extractelement <2 x i1> [[TMP28]], i32 0
-; TFA_INTERLEAVE-NEXT: br i1 [[TMP30]], label [[END:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; TFA_INTERLEAVE-NEXT: [[TMP26:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; TFA_INTERLEAVE-NEXT: [[TMP28:%.*]] = xor i1 [[TMP26]], true
+; TFA_INTERLEAVE-NEXT: br i1 [[TMP28]], label [[END:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; TFA_INTERLEAVE: end:
; TFA_INTERLEAVE-NEXT: ret void
;
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/masked-call.ll b/llvm/test/Transforms/LoopVectorize/AArch64/masked-call.ll
index 6029095bbe7b1..c4c4c9c45c96b 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/masked-call.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/masked-call.ll
@@ -66,8 +66,8 @@ define void @test_widen(ptr noalias %a, ptr readnone %b) #4 {
; TFCOMMON-NEXT: call void @llvm.masked.store.nxv2i64.p0(<vscale x 2 x i64> [[TMP6]], ptr [[TMP7]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; TFCOMMON-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP4]]
; TFCOMMON-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX_NEXT]], i64 1025)
-; TFCOMMON-NEXT: [[TMP8:%.*]] = xor <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
-; TFCOMMON-NEXT: [[TMP9:%.*]] = extractelement <vscale x 2 x i1> [[TMP8]], i32 0
+; TFCOMMON-NEXT: [[TMP8:%.*]] = extractelement <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; TFCOMMON-NEXT: [[TMP9:%.*]] = xor i1 [[TMP8]], true
; TFCOMMON-NEXT: br i1 [[TMP9]], label %[[FOR_COND_CLEANUP:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; TFCOMMON: [[FOR_COND_CLEANUP]]:
; TFCOMMON-NEXT: ret void
@@ -106,8 +106,8 @@ define void @test_widen(ptr noalias %a, ptr readnone %b) #4 {
; TFA_INTERLEAVE-NEXT: [[TMP19:%.*]] = add i64 [[INDEX_NEXT]], [[TMP18]]
; TFA_INTERLEAVE-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX_NEXT]], i64 1025)
; TFA_INTERLEAVE-NEXT: [[ACTIVE_LANE_MASK_NEXT4]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[TMP19]], i64 1025)
-; TFA_INTERLEAVE-NEXT: [[TMP20:%.*]] = xor <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
-; TFA_INTERLEAVE-NEXT: [[TMP21:%.*]] = extractelement <vscale x 2 x i1> [[TMP20]], i32 0
+; TFA_INTERLEAVE-NEXT: [[TMP20:%.*]] = extractelement <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i32 0
+; TFA_INTERLEAVE-NEXT: [[TMP21:%.*]] = xor i1 [[TMP20]], true
; TFA_INTERLEAVE-NEXT: br i1 [[TMP21]], label %[[FOR_COND_CLEANUP:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; TFA_INTERLEAVE: [[FOR_COND_CLEANUP]]:
; TFA_INTERLEAVE-NEXT: ret void
@@ -200,8 +200,8 @@ define void @test_if_then(ptr noalias %a, ptr readnone %b) #4 {
; TFCOMMON-NEXT: call void @llvm.masked.store.nxv2i64.p0(<vscale x 2 x i64> [[PREDPHI]], ptr [[TMP9]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; TFCOMMON-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP4]]
; TFCOMMON-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX_NEXT]], i64 1025)
-; TFCOMMON-NEXT: [[TMP10:%.*]] = xor <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
-; TFCOMMON-NEXT: [[TMP11:%.*]] = extractelement <vscal...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/152628
More information about the llvm-commits
mailing list