Skip to content

Commit bd05376

Browse files
alexey-bataevalexey-bataev
committed
[SLP]Improve multinode analysis.
Changes the preliminary multinode analysis: 1. Introduced scores for reversed loads/extractelements. 2. Improved shallow score calculation. 3. Lowered the cost of external uses (no need to consider it several times, just ones). 4. The initial lane for analysis is the one with the minimal possible reorderings. These changes in general shall reduce compile time and improve the reordering in many cases. Part of D57059. Differential Revision: https://reviews.llvm.org/D101109
1 parent 135d5d4 commit bd05376

16 files changed

+328
-204
lines changed

‎llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

Lines changed: 197 additions & 56 deletions
Large diffs are not rendered by default.

‎llvm/test/Transforms/SLPVectorizer/AArch64/transpose-inseltpoison.ll

Lines changed: 11 additions & 19 deletions
Original file line numberDiff line numberDiff line change
@@ -167,25 +167,17 @@ define <4 x i32> @build_vec_v4i32_reuse_1(<2 x i32> %v0, <2 x i32> %v1) {
167167

168168
define <4 x i32> @build_vec_v4i32_3_binops(<2 x i32> %v0, <2 x i32> %v1) {
169169
; CHECK-LABEL: @build_vec_v4i32_3_binops(
170-
; CHECK-NEXT: [[V0_0:%.*]] = extractelement <2 x i32> [[V0:%.*]], i64 0
171-
; CHECK-NEXT: [[V0_1:%.*]] = extractelement <2 x i32> [[V0]], i64 1
172-
; CHECK-NEXT: [[V1_0:%.*]] = extractelement <2 x i32> [[V1:%.*]], i64 0
173-
; CHECK-NEXT: [[V1_1:%.*]] = extractelement <2 x i32> [[V1]], i64 1
174-
; CHECK-NEXT: [[TMP0_0:%.*]] = add i32 [[V0_0]], [[V1_0]]
175-
; CHECK-NEXT: [[TMP0_1:%.*]] = add i32 [[V0_1]], [[V1_1]]
176-
; CHECK-NEXT: [[TMP1_0:%.*]] = mul i32 [[V0_0]], [[V1_0]]
177-
; CHECK-NEXT: [[TMP1_1:%.*]] = mul i32 [[V0_1]], [[V1_1]]
178-
; CHECK-NEXT: [[TMP1:%.*]] = xor <2 x i32> [[V0]], [[V1]]
179-
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> poison, <2 x i32> zeroinitializer
180-
; CHECK-NEXT: [[TMP3:%.*]] = xor <2 x i32> [[V0]], [[V1]]
181-
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <2 x i32> [[TMP3]], <2 x i32> poison, <2 x i32> <i32 1, i32 1>
182-
; CHECK-NEXT: [[TMP2_0:%.*]] = add i32 [[TMP0_0]], [[TMP0_1]]
183-
; CHECK-NEXT: [[TMP2_1:%.*]] = add i32 [[TMP1_0]], [[TMP1_1]]
184-
; CHECK-NEXT: [[TMP5:%.*]] = add <2 x i32> [[TMP2]], [[TMP4]]
185-
; CHECK-NEXT: [[TMP3_0:%.*]] = insertelement <4 x i32> poison, i32 [[TMP2_0]], i64 0
186-
; CHECK-NEXT: [[TMP3_1:%.*]] = insertelement <4 x i32> [[TMP3_0]], i32 [[TMP2_1]], i64 1
187-
; CHECK-NEXT: [[TMP6:%.*]] = shufflevector <2 x i32> [[TMP5]], <2 x i32> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
188-
; CHECK-NEXT: [[TMP3_31:%.*]] = shufflevector <4 x i32> [[TMP3_1]], <4 x i32> [[TMP6]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
170+
; CHECK-NEXT: [[TMP1:%.*]] = add <2 x i32> [[V0:%.*]], [[V1:%.*]]
171+
; CHECK-NEXT: [[TMP2:%.*]] = mul <2 x i32> [[V0]], [[V1]]
172+
; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP2]], <2 x i32> <i32 1, i32 2>
173+
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP2]], <2 x i32> <i32 0, i32 3>
174+
; CHECK-NEXT: [[TMP5:%.*]] = xor <2 x i32> [[V0]], [[V1]]
175+
; CHECK-NEXT: [[TMP6:%.*]] = shufflevector <2 x i32> [[TMP5]], <2 x i32> poison, <2 x i32> zeroinitializer
176+
; CHECK-NEXT: [[TMP7:%.*]] = xor <2 x i32> [[V0]], [[V1]]
177+
; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <2 x i32> [[TMP7]], <2 x i32> poison, <2 x i32> <i32 1, i32 1>
178+
; CHECK-NEXT: [[TMP9:%.*]] = add <2 x i32> [[TMP4]], [[TMP3]]
179+
; CHECK-NEXT: [[TMP10:%.*]] = add <2 x i32> [[TMP6]], [[TMP8]]
180+
; CHECK-NEXT: [[TMP3_31:%.*]] = shufflevector <2 x i32> [[TMP9]], <2 x i32> [[TMP10]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
189181
; CHECK-NEXT: ret <4 x i32> [[TMP3_31]]
190182
;
191183
%v0.0 = extractelement <2 x i32> %v0, i32 0

‎llvm/test/Transforms/SLPVectorizer/AArch64/transpose.ll

Lines changed: 11 additions & 19 deletions
Original file line numberDiff line numberDiff line change
@@ -167,25 +167,17 @@ define <4 x i32> @build_vec_v4i32_reuse_1(<2 x i32> %v0, <2 x i32> %v1) {
167167

168168
define <4 x i32> @build_vec_v4i32_3_binops(<2 x i32> %v0, <2 x i32> %v1) {
169169
; CHECK-LABEL: @build_vec_v4i32_3_binops(
170-
; CHECK-NEXT: [[V0_0:%.*]] = extractelement <2 x i32> [[V0:%.*]], i64 0
171-
; CHECK-NEXT: [[V0_1:%.*]] = extractelement <2 x i32> [[V0]], i64 1
172-
; CHECK-NEXT: [[V1_0:%.*]] = extractelement <2 x i32> [[V1:%.*]], i64 0
173-
; CHECK-NEXT: [[V1_1:%.*]] = extractelement <2 x i32> [[V1]], i64 1
174-
; CHECK-NEXT: [[TMP0_0:%.*]] = add i32 [[V0_0]], [[V1_0]]
175-
; CHECK-NEXT: [[TMP0_1:%.*]] = add i32 [[V0_1]], [[V1_1]]
176-
; CHECK-NEXT: [[TMP1_0:%.*]] = mul i32 [[V0_0]], [[V1_0]]
177-
; CHECK-NEXT: [[TMP1_1:%.*]] = mul i32 [[V0_1]], [[V1_1]]
178-
; CHECK-NEXT: [[TMP1:%.*]] = xor <2 x i32> [[V0]], [[V1]]
179-
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> poison, <2 x i32> zeroinitializer
180-
; CHECK-NEXT: [[TMP3:%.*]] = xor <2 x i32> [[V0]], [[V1]]
181-
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <2 x i32> [[TMP3]], <2 x i32> poison, <2 x i32> <i32 1, i32 1>
182-
; CHECK-NEXT: [[TMP2_0:%.*]] = add i32 [[TMP0_0]], [[TMP0_1]]
183-
; CHECK-NEXT: [[TMP2_1:%.*]] = add i32 [[TMP1_0]], [[TMP1_1]]
184-
; CHECK-NEXT: [[TMP5:%.*]] = add <2 x i32> [[TMP2]], [[TMP4]]
185-
; CHECK-NEXT: [[TMP3_0:%.*]] = insertelement <4 x i32> undef, i32 [[TMP2_0]], i64 0
186-
; CHECK-NEXT: [[TMP3_1:%.*]] = insertelement <4 x i32> [[TMP3_0]], i32 [[TMP2_1]], i64 1
187-
; CHECK-NEXT: [[TMP6:%.*]] = shufflevector <2 x i32> [[TMP5]], <2 x i32> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
188-
; CHECK-NEXT: [[TMP3_31:%.*]] = shufflevector <4 x i32> [[TMP3_1]], <4 x i32> [[TMP6]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
170+
; CHECK-NEXT: [[TMP1:%.*]] = add <2 x i32> [[V0:%.*]], [[V1:%.*]]
171+
; CHECK-NEXT: [[TMP2:%.*]] = mul <2 x i32> [[V0]], [[V1]]
172+
; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP2]], <2 x i32> <i32 1, i32 2>
173+
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP2]], <2 x i32> <i32 0, i32 3>
174+
; CHECK-NEXT: [[TMP5:%.*]] = xor <2 x i32> [[V0]], [[V1]]
175+
; CHECK-NEXT: [[TMP6:%.*]] = shufflevector <2 x i32> [[TMP5]], <2 x i32> poison, <2 x i32> zeroinitializer
176+
; CHECK-NEXT: [[TMP7:%.*]] = xor <2 x i32> [[V0]], [[V1]]
177+
; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <2 x i32> [[TMP7]], <2 x i32> poison, <2 x i32> <i32 1, i32 1>
178+
; CHECK-NEXT: [[TMP9:%.*]] = add <2 x i32> [[TMP4]], [[TMP3]]
179+
; CHECK-NEXT: [[TMP10:%.*]] = add <2 x i32> [[TMP6]], [[TMP8]]
180+
; CHECK-NEXT: [[TMP3_31:%.*]] = shufflevector <2 x i32> [[TMP9]], <2 x i32> [[TMP10]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
189181
; CHECK-NEXT: ret <4 x i32> [[TMP3_31]]
190182
;
191183
%v0.0 = extractelement <2 x i32> %v0, i32 0

‎llvm/test/Transforms/SLPVectorizer/AArch64/vectorize-free-extracts-inserts.ll

Lines changed: 11 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -282,19 +282,21 @@ define void @extracts_jumbled_4_lanes(<9 x double>* %ptr.1, <4 x double>* %ptr.2
282282
; CHECK-NEXT: [[V2_LANE_0:%.*]] = extractelement <4 x double> [[V_2]], i32 0
283283
; CHECK-NEXT: [[V2_LANE_1:%.*]] = extractelement <4 x double> [[V_2]], i32 1
284284
; CHECK-NEXT: [[V2_LANE_2:%.*]] = extractelement <4 x double> [[V_2]], i32 2
285-
; CHECK-NEXT: [[A_LANE_0:%.*]] = fmul double [[V1_LANE_0]], [[V2_LANE_2]]
286-
; CHECK-NEXT: [[A_LANE_1:%.*]] = fmul double [[V1_LANE_2]], [[V2_LANE_1]]
287-
; CHECK-NEXT: [[A_LANE_2:%.*]] = fmul double [[V1_LANE_1]], [[V2_LANE_2]]
288-
; CHECK-NEXT: [[A_LANE_3:%.*]] = fmul double [[V1_LANE_3]], [[V2_LANE_0]]
289-
; CHECK-NEXT: [[A_INS_0:%.*]] = insertelement <9 x double> undef, double [[A_LANE_0]], i32 0
290-
; CHECK-NEXT: [[A_INS_1:%.*]] = insertelement <9 x double> [[A_INS_0]], double [[A_LANE_1]], i32 1
291-
; CHECK-NEXT: [[A_INS_2:%.*]] = insertelement <9 x double> [[A_INS_1]], double [[A_LANE_2]], i32 2
292-
; CHECK-NEXT: [[A_INS_3:%.*]] = insertelement <9 x double> [[A_INS_2]], double [[A_LANE_3]], i32 3
285+
; CHECK-NEXT: [[TMP0:%.*]] = insertelement <4 x double> poison, double [[V1_LANE_0]], i32 0
286+
; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x double> [[TMP0]], double [[V1_LANE_2]], i32 1
287+
; CHECK-NEXT: [[TMP2:%.*]] = insertelement <4 x double> [[TMP1]], double [[V1_LANE_1]], i32 2
288+
; CHECK-NEXT: [[TMP3:%.*]] = insertelement <4 x double> [[TMP2]], double [[V1_LANE_3]], i32 3
289+
; CHECK-NEXT: [[TMP4:%.*]] = insertelement <4 x double> poison, double [[V2_LANE_2]], i32 0
290+
; CHECK-NEXT: [[TMP5:%.*]] = insertelement <4 x double> [[TMP4]], double [[V2_LANE_1]], i32 1
291+
; CHECK-NEXT: [[TMP6:%.*]] = insertelement <4 x double> [[TMP5]], double [[V2_LANE_2]], i32 2
292+
; CHECK-NEXT: [[TMP7:%.*]] = insertelement <4 x double> [[TMP6]], double [[V2_LANE_0]], i32 3
293+
; CHECK-NEXT: [[TMP8:%.*]] = fmul <4 x double> [[TMP3]], [[TMP7]]
294+
; CHECK-NEXT: [[TMP9:%.*]] = shufflevector <4 x double> [[TMP8]], <4 x double> poison, <9 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
293295
; CHECK-NEXT: call void @use(double [[V1_LANE_0]])
294296
; CHECK-NEXT: call void @use(double [[V1_LANE_1]])
295297
; CHECK-NEXT: call void @use(double [[V1_LANE_2]])
296298
; CHECK-NEXT: call void @use(double [[V1_LANE_3]])
297-
; CHECK-NEXT: store <9 x double> [[A_INS_3]], <9 x double>* [[PTR_1]], align 8
299+
; CHECK-NEXT: store <9 x double> [[TMP9]], <9 x double>* [[PTR_1]], align 8
298300
; CHECK-NEXT: ret void
299301
;
300302
bb:

‎llvm/test/Transforms/SLPVectorizer/X86/PR39774.ll

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
22
; RUN: opt -slp-vectorizer -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=skylake -slp-threshold=-6 | FileCheck %s --check-prefix=CHECK
3-
; RUN: opt -slp-vectorizer -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=skylake -slp-threshold=-8 -slp-min-tree-size=6 | FileCheck %s --check-prefix=FORCE_REDUCTION
3+
; RUN: opt -slp-vectorizer -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=skylake -slp-threshold=-7 -slp-min-tree-size=6 | FileCheck %s --check-prefix=FORCE_REDUCTION
44

55
define void @Test(i32) {
66
; CHECK-LABEL: @Test(

‎llvm/test/Transforms/SLPVectorizer/X86/addsub.ll

Lines changed: 12 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -342,18 +342,18 @@ define void @vec_shuff_reorder() #0 {
342342
; CHECK-LABEL: @vec_shuff_reorder(
343343
; CHECK-NEXT: [[TMP1:%.*]] = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 0), align 4
344344
; CHECK-NEXT: [[TMP2:%.*]] = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 0), align 4
345-
; CHECK-NEXT: [[TMP3:%.*]] = load <2 x float>, <2 x float>* bitcast (float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 1) to <2 x float>*), align 4
346-
; CHECK-NEXT: [[TMP4:%.*]] = load <2 x float>, <2 x float>* bitcast (float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 1) to <2 x float>*), align 4
347-
; CHECK-NEXT: [[TMP5:%.*]] = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 3), align 4
348-
; CHECK-NEXT: [[TMP6:%.*]] = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 3), align 4
349-
; CHECK-NEXT: [[TMP7:%.*]] = insertelement <4 x float> poison, float [[TMP2]], i32 0
350-
; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <2 x float> [[TMP3]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
351-
; CHECK-NEXT: [[TMP9:%.*]] = shufflevector <4 x float> [[TMP7]], <4 x float> [[TMP8]], <4 x i32> <i32 0, i32 4, i32 5, i32 3>
352-
; CHECK-NEXT: [[TMP10:%.*]] = insertelement <4 x float> [[TMP9]], float [[TMP5]], i32 3
353-
; CHECK-NEXT: [[TMP11:%.*]] = insertelement <4 x float> poison, float [[TMP1]], i32 0
354-
; CHECK-NEXT: [[TMP12:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
355-
; CHECK-NEXT: [[TMP13:%.*]] = shufflevector <4 x float> [[TMP11]], <4 x float> [[TMP12]], <4 x i32> <i32 0, i32 4, i32 5, i32 3>
356-
; CHECK-NEXT: [[TMP14:%.*]] = insertelement <4 x float> [[TMP13]], float [[TMP6]], i32 3
345+
; CHECK-NEXT: [[TMP3:%.*]] = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 1), align 4
346+
; CHECK-NEXT: [[TMP4:%.*]] = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 1), align 4
347+
; CHECK-NEXT: [[TMP5:%.*]] = load <2 x float>, <2 x float>* bitcast (float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 2) to <2 x float>*), align 4
348+
; CHECK-NEXT: [[TMP6:%.*]] = load <2 x float>, <2 x float>* bitcast (float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 2) to <2 x float>*), align 4
349+
; CHECK-NEXT: [[TMP7:%.*]] = insertelement <4 x float> poison, float [[TMP1]], i32 0
350+
; CHECK-NEXT: [[TMP8:%.*]] = insertelement <4 x float> [[TMP7]], float [[TMP3]], i32 1
351+
; CHECK-NEXT: [[TMP9:%.*]] = shufflevector <2 x float> [[TMP5]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
352+
; CHECK-NEXT: [[TMP10:%.*]] = shufflevector <4 x float> [[TMP8]], <4 x float> [[TMP9]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
353+
; CHECK-NEXT: [[TMP11:%.*]] = insertelement <4 x float> poison, float [[TMP2]], i32 0
354+
; CHECK-NEXT: [[TMP12:%.*]] = insertelement <4 x float> [[TMP11]], float [[TMP4]], i32 1
355+
; CHECK-NEXT: [[TMP13:%.*]] = shufflevector <2 x float> [[TMP6]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
356+
; CHECK-NEXT: [[TMP14:%.*]] = shufflevector <4 x float> [[TMP12]], <4 x float> [[TMP13]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
357357
; CHECK-NEXT: [[TMP15:%.*]] = fadd <4 x float> [[TMP10]], [[TMP14]]
358358
; CHECK-NEXT: [[TMP16:%.*]] = fsub <4 x float> [[TMP10]], [[TMP14]]
359359
; CHECK-NEXT: [[TMP17:%.*]] = shufflevector <4 x float> [[TMP15]], <4 x float> [[TMP16]], <4 x i32> <i32 0, i32 5, i32 2, i32 7>

‎llvm/test/Transforms/SLPVectorizer/X86/commutativity.ll

Lines changed: 10 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -16,21 +16,21 @@
1616

1717
define void @splat(i8 %a, i8 %b, i8 %c) {
1818
; SSE-LABEL: @splat(
19-
; SSE-NEXT: [[TMP1:%.*]] = insertelement <16 x i8> poison, i8 [[C:%.*]], i32 0
20-
; SSE-NEXT: [[SHUFFLE:%.*]] = shufflevector <16 x i8> [[TMP1]], <16 x i8> poison, <16 x i32> zeroinitializer
21-
; SSE-NEXT: [[TMP2:%.*]] = insertelement <16 x i8> poison, i8 [[A:%.*]], i32 0
22-
; SSE-NEXT: [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[B:%.*]], i32 1
23-
; SSE-NEXT: [[SHUFFLE1:%.*]] = shufflevector <16 x i8> [[TMP3]], <16 x i8> poison, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
19+
; SSE-NEXT: [[TMP1:%.*]] = insertelement <16 x i8> poison, i8 [[A:%.*]], i32 0
20+
; SSE-NEXT: [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[B:%.*]], i32 1
21+
; SSE-NEXT: [[SHUFFLE:%.*]] = shufflevector <16 x i8> [[TMP2]], <16 x i8> poison, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
22+
; SSE-NEXT: [[TMP3:%.*]] = insertelement <16 x i8> poison, i8 [[C:%.*]], i32 0
23+
; SSE-NEXT: [[SHUFFLE1:%.*]] = shufflevector <16 x i8> [[TMP3]], <16 x i8> poison, <16 x i32> zeroinitializer
2424
; SSE-NEXT: [[TMP4:%.*]] = xor <16 x i8> [[SHUFFLE]], [[SHUFFLE1]]
2525
; SSE-NEXT: store <16 x i8> [[TMP4]], <16 x i8>* bitcast ([32 x i8]* @cle to <16 x i8>*), align 16
2626
; SSE-NEXT: ret void
2727
;
2828
; AVX-LABEL: @splat(
29-
; AVX-NEXT: [[TMP1:%.*]] = insertelement <16 x i8> poison, i8 [[C:%.*]], i32 0
30-
; AVX-NEXT: [[SHUFFLE:%.*]] = shufflevector <16 x i8> [[TMP1]], <16 x i8> poison, <16 x i32> zeroinitializer
31-
; AVX-NEXT: [[TMP2:%.*]] = insertelement <16 x i8> poison, i8 [[A:%.*]], i32 0
32-
; AVX-NEXT: [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[B:%.*]], i32 1
33-
; AVX-NEXT: [[SHUFFLE1:%.*]] = shufflevector <16 x i8> [[TMP3]], <16 x i8> poison, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
29+
; AVX-NEXT: [[TMP1:%.*]] = insertelement <16 x i8> poison, i8 [[A:%.*]], i32 0
30+
; AVX-NEXT: [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[B:%.*]], i32 1
31+
; AVX-NEXT: [[SHUFFLE:%.*]] = shufflevector <16 x i8> [[TMP2]], <16 x i8> poison, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
32+
; AVX-NEXT: [[TMP3:%.*]] = insertelement <16 x i8> poison, i8 [[C:%.*]], i32 0
33+
; AVX-NEXT: [[SHUFFLE1:%.*]] = shufflevector <16 x i8> [[TMP3]], <16 x i8> poison, <16 x i32> zeroinitializer
3434
; AVX-NEXT: [[TMP4:%.*]] = xor <16 x i8> [[SHUFFLE]], [[SHUFFLE1]]
3535
; AVX-NEXT: store <16 x i8> [[TMP4]], <16 x i8>* bitcast ([32 x i8]* @cle to <16 x i8>*), align 16
3636
; AVX-NEXT: ret void

‎llvm/test/Transforms/SLPVectorizer/X86/crash_exceed_scheduling.ll

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -34,9 +34,9 @@ define void @exceed(double %0, double %1) {
3434
; CHECK-NEXT: [[TMP11:%.*]] = fadd fast <2 x double> [[TMP3]], [[TMP5]]
3535
; CHECK-NEXT: [[TMP12:%.*]] = fmul fast <2 x double> [[TMP10]], [[TMP11]]
3636
; CHECK-NEXT: [[IXX101:%.*]] = fsub double undef, undef
37-
; CHECK-NEXT: [[TMP13:%.*]] = insertelement <2 x double> <double poison, double undef>, double [[TMP7]], i32 0
38-
; CHECK-NEXT: [[TMP14:%.*]] = insertelement <2 x double> <double undef, double poison>, double [[TMP1]], i32 1
39-
; CHECK-NEXT: [[TMP15:%.*]] = fmul fast <2 x double> [[TMP13]], [[TMP14]]
37+
; CHECK-NEXT: [[TMP13:%.*]] = insertelement <2 x double> poison, double [[TMP1]], i32 1
38+
; CHECK-NEXT: [[TMP14:%.*]] = insertelement <2 x double> [[TMP13]], double [[TMP7]], i32 0
39+
; CHECK-NEXT: [[TMP15:%.*]] = fmul fast <2 x double> [[TMP14]], undef
4040
; CHECK-NEXT: switch i32 undef, label [[BB1:%.*]] [
4141
; CHECK-NEXT: i32 0, label [[BB2:%.*]]
4242
; CHECK-NEXT: ]

‎llvm/test/Transforms/SLPVectorizer/X86/crash_smallpt.ll

Lines changed: 11 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -30,15 +30,19 @@ define void @main() #0 {
3030
; CHECK-NEXT: br i1 undef, label [[COND_TRUE63_US:%.*]], label [[COND_FALSE66_US:%.*]]
3131
; CHECK: cond.false66.us:
3232
; CHECK-NEXT: [[ADD_I276_US:%.*]] = fadd double 0.000000e+00, undef
33-
; CHECK-NEXT: [[TMP0:%.*]] = insertelement <2 x double> <double poison, double undef>, double [[ADD_I276_US]], i32 0
34-
; CHECK-NEXT: [[TMP1:%.*]] = fadd <2 x double> [[TMP0]], <double 0.000000e+00, double 0xBFA5CC2D1960285F>
33+
; CHECK-NEXT: [[TMP0:%.*]] = insertelement <2 x double> <double poison, double 0xBFA5CC2D1960285F>, double [[ADD_I276_US]], i32 0
34+
; CHECK-NEXT: [[TMP1:%.*]] = fadd <2 x double> <double 0.000000e+00, double undef>, [[TMP0]]
3535
; CHECK-NEXT: [[TMP2:%.*]] = fmul <2 x double> [[TMP1]], <double 1.400000e+02, double 1.400000e+02>
3636
; CHECK-NEXT: [[TMP3:%.*]] = fadd <2 x double> [[TMP2]], <double 5.000000e+01, double 5.200000e+01>
37-
; CHECK-NEXT: [[TMP4:%.*]] = fmul <2 x double> undef, [[TMP1]]
38-
; CHECK-NEXT: [[TMP5:%.*]] = bitcast double* [[AGG_TMP99208_SROA_0_0_IDX]] to <2 x double>*
39-
; CHECK-NEXT: store <2 x double> [[TMP3]], <2 x double>* [[TMP5]], align 8
40-
; CHECK-NEXT: [[TMP6:%.*]] = bitcast double* [[AGG_TMP101211_SROA_0_0_IDX]] to <2 x double>*
41-
; CHECK-NEXT: store <2 x double> [[TMP4]], <2 x double>* [[TMP6]], align 8
37+
; CHECK-NEXT: [[TMP4:%.*]] = extractelement <2 x double> [[TMP1]], i32 0
38+
; CHECK-NEXT: [[TMP5:%.*]] = extractelement <2 x double> [[TMP1]], i32 1
39+
; CHECK-NEXT: [[TMP6:%.*]] = insertelement <2 x double> <double poison, double undef>, double [[TMP4]], i32 0
40+
; CHECK-NEXT: [[TMP7:%.*]] = insertelement <2 x double> <double undef, double poison>, double [[TMP5]], i32 1
41+
; CHECK-NEXT: [[TMP8:%.*]] = fmul <2 x double> [[TMP6]], [[TMP7]]
42+
; CHECK-NEXT: [[TMP9:%.*]] = bitcast double* [[AGG_TMP99208_SROA_0_0_IDX]] to <2 x double>*
43+
; CHECK-NEXT: store <2 x double> [[TMP3]], <2 x double>* [[TMP9]], align 8
44+
; CHECK-NEXT: [[TMP10:%.*]] = bitcast double* [[AGG_TMP101211_SROA_0_0_IDX]] to <2 x double>*
45+
; CHECK-NEXT: store <2 x double> [[TMP8]], <2 x double>* [[TMP10]], align 8
4246
; CHECK-NEXT: unreachable
4347
; CHECK: cond.true63.us:
4448
; CHECK-NEXT: unreachable

‎llvm/test/Transforms/SLPVectorizer/X86/extractelement.ll

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -85,7 +85,7 @@ define float @f_used_twice_in_tree(<2 x float> %x) {
8585
; THRESH1-NEXT: [[TMP1:%.*]] = extractelement <2 x float> [[X:%.*]], i32 1
8686
; THRESH1-NEXT: [[TMP2:%.*]] = insertelement <2 x float> poison, float [[TMP1]], i32 0
8787
; THRESH1-NEXT: [[TMP3:%.*]] = insertelement <2 x float> [[TMP2]], float [[TMP1]], i32 1
88-
; THRESH1-NEXT: [[TMP4:%.*]] = fmul <2 x float> [[X]], [[TMP3]]
88+
; THRESH1-NEXT: [[TMP4:%.*]] = fmul <2 x float> [[TMP3]], [[X]]
8989
; THRESH1-NEXT: [[TMP5:%.*]] = extractelement <2 x float> [[TMP4]], i32 0
9090
; THRESH1-NEXT: [[TMP6:%.*]] = extractelement <2 x float> [[TMP4]], i32 1
9191
; THRESH1-NEXT: [[ADD:%.*]] = fadd float [[TMP5]], [[TMP6]]
@@ -95,7 +95,7 @@ define float @f_used_twice_in_tree(<2 x float> %x) {
9595
; THRESH2-NEXT: [[TMP1:%.*]] = extractelement <2 x float> [[X:%.*]], i32 1
9696
; THRESH2-NEXT: [[TMP2:%.*]] = insertelement <2 x float> poison, float [[TMP1]], i32 0
9797
; THRESH2-NEXT: [[TMP3:%.*]] = insertelement <2 x float> [[TMP2]], float [[TMP1]], i32 1
98-
; THRESH2-NEXT: [[TMP4:%.*]] = fmul <2 x float> [[X]], [[TMP3]]
98+
; THRESH2-NEXT: [[TMP4:%.*]] = fmul <2 x float> [[TMP3]], [[X]]
9999
; THRESH2-NEXT: [[TMP5:%.*]] = extractelement <2 x float> [[TMP4]], i32 0
100100
; THRESH2-NEXT: [[TMP6:%.*]] = extractelement <2 x float> [[TMP4]], i32 1
101101
; THRESH2-NEXT: [[ADD:%.*]] = fadd float [[TMP5]], [[TMP6]]

0 commit comments

Comments
 (0)