Skip to content

Commit 59fce6b

Browse files
dobbelaj-snpspratlucas
dobbelaj-snps
authored and
pratlucas
committed
[NFC] make clang/test/CodeGen/arm_neon_intrinsics.c resistent to function attribute id changes
When introducing support for @llvm.experimental.noalias.scope.decl, this tests started failing because it checks (for no good reason) for a function attribute id of '#8' which now becomes '#9' Reviewed By: pratlucas Differential Revision: https://reviews.llvm.org/D94233
1 parent 82f5ee3 commit 59fce6b

File tree

1 file changed

+30
-30
lines changed

1 file changed

+30
-30
lines changed

‎clang/test/CodeGen/arm_neon_intrinsics.c‎

Lines changed: 30 additions & 30 deletions
Original file line numberDiff line numberDiff line change
@@ -7114,7 +7114,7 @@ uint64x2_t test_vmlal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c) {
71147114
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
71157115
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[B:%.*]] to <8 x i8>
71167116
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
7117-
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]]) #8
7117+
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]])
71187118
// CHECK: [[ADD:%.*]] = add <4 x i32> [[A:%.*]], [[VMULL2_I]]
71197119
// CHECK: ret <4 x i32> [[ADD]]
71207120
int32x4_t test_vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t c) {
@@ -7127,7 +7127,7 @@ int32x4_t test_vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t c) {
71277127
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
71287128
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[B:%.*]] to <8 x i8>
71297129
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
7130-
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmulls.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]]) #8
7130+
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmulls.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]])
71317131
// CHECK: [[ADD:%.*]] = add <2 x i64> [[A:%.*]], [[VMULL2_I]]
71327132
// CHECK: ret <2 x i64> [[ADD]]
71337133
int64x2_t test_vmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t c) {
@@ -7140,7 +7140,7 @@ int64x2_t test_vmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t c) {
71407140
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
71417141
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[B:%.*]] to <8 x i8>
71427142
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
7143-
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]]) #8
7143+
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]])
71447144
// CHECK: [[ADD:%.*]] = add <4 x i32> [[A:%.*]], [[VMULL2_I]]
71457145
// CHECK: ret <4 x i32> [[ADD]]
71467146
uint32x4_t test_vmlal_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c) {
@@ -7153,7 +7153,7 @@ uint32x4_t test_vmlal_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c) {
71537153
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
71547154
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[B:%.*]] to <8 x i8>
71557155
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
7156-
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmullu.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]]) #8
7156+
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmullu.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]])
71577157
// CHECK: [[ADD:%.*]] = add <2 x i64> [[A:%.*]], [[VMULL2_I]]
71587158
// CHECK: ret <2 x i64> [[ADD]]
71597159
uint64x2_t test_vmlal_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c) {
@@ -7618,7 +7618,7 @@ uint64x2_t test_vmlsl_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c) {
76187618
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
76197619
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[B:%.*]] to <8 x i8>
76207620
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
7621-
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]]) #8
7621+
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]])
76227622
// CHECK: [[SUB:%.*]] = sub <4 x i32> [[A:%.*]], [[VMULL2_I]]
76237623
// CHECK: ret <4 x i32> [[SUB]]
76247624
int32x4_t test_vmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t c) {
@@ -7631,7 +7631,7 @@ int32x4_t test_vmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t c) {
76317631
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
76327632
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[B:%.*]] to <8 x i8>
76337633
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
7634-
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmulls.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]]) #8
7634+
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmulls.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]])
76357635
// CHECK: [[SUB:%.*]] = sub <2 x i64> [[A:%.*]], [[VMULL2_I]]
76367636
// CHECK: ret <2 x i64> [[SUB]]
76377637
int64x2_t test_vmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t c) {
@@ -7644,7 +7644,7 @@ int64x2_t test_vmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t c) {
76447644
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
76457645
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[B:%.*]] to <8 x i8>
76467646
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
7647-
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]]) #8
7647+
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]])
76487648
// CHECK: [[SUB:%.*]] = sub <4 x i32> [[A:%.*]], [[VMULL2_I]]
76497649
// CHECK: ret <4 x i32> [[SUB]]
76507650
uint32x4_t test_vmlsl_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c) {
@@ -7657,7 +7657,7 @@ uint32x4_t test_vmlsl_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c) {
76577657
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
76587658
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[B:%.*]] to <8 x i8>
76597659
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
7660-
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmullu.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]]) #8
7660+
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmullu.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]])
76617661
// CHECK: [[SUB:%.*]] = sub <2 x i64> [[A:%.*]], [[VMULL2_I]]
76627662
// CHECK: ret <2 x i64> [[SUB]]
76637663
uint64x2_t test_vmlsl_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c) {
@@ -8497,7 +8497,7 @@ poly16x8_t test_vmull_p8(poly8x8_t a, poly8x8_t b) {
84978497
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
84988498
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[A:%.*]] to <8 x i8>
84998499
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
8500-
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> [[A]], <4 x i16> [[LANE]]) #8
8500+
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> [[A]], <4 x i16> [[LANE]])
85018501
// CHECK: ret <4 x i32> [[VMULL2_I]]
85028502
int32x4_t test_vmull_lane_s16(int16x4_t a, int16x4_t b) {
85038503
return vmull_lane_s16(a, b, 3);
@@ -8509,7 +8509,7 @@ int32x4_t test_vmull_lane_s16(int16x4_t a, int16x4_t b) {
85098509
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
85108510
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[A:%.*]] to <8 x i8>
85118511
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
8512-
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmulls.v2i64(<2 x i32> [[A]], <2 x i32> [[LANE]]) #8
8512+
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmulls.v2i64(<2 x i32> [[A]], <2 x i32> [[LANE]])
85138513
// CHECK: ret <2 x i64> [[VMULL2_I]]
85148514
int64x2_t test_vmull_lane_s32(int32x2_t a, int32x2_t b) {
85158515
return vmull_lane_s32(a, b, 1);
@@ -8521,7 +8521,7 @@ int64x2_t test_vmull_lane_s32(int32x2_t a, int32x2_t b) {
85218521
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
85228522
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[A:%.*]] to <8 x i8>
85238523
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
8524-
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16> [[A]], <4 x i16> [[LANE]]) #8
8524+
// CHECK: [[VMULL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16> [[A]], <4 x i16> [[LANE]])
85258525
// CHECK: ret <4 x i32> [[VMULL2_I]]
85268526
uint32x4_t test_vmull_lane_u16(uint16x4_t a, uint16x4_t b) {
85278527
return vmull_lane_u16(a, b, 3);
@@ -8533,7 +8533,7 @@ uint32x4_t test_vmull_lane_u16(uint16x4_t a, uint16x4_t b) {
85338533
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
85348534
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[A:%.*]] to <8 x i8>
85358535
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
8536-
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmullu.v2i64(<2 x i32> [[A]], <2 x i32> [[LANE]]) #8
8536+
// CHECK: [[VMULL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vmullu.v2i64(<2 x i32> [[A]], <2 x i32> [[LANE]])
85378537
// CHECK: ret <2 x i64> [[VMULL2_I]]
85388538
uint64x2_t test_vmull_lane_u32(uint32x2_t a, uint32x2_t b) {
85398539
return vmull_lane_u32(a, b, 1);
@@ -9822,8 +9822,8 @@ int64x2_t test_vqdmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c) {
98229822
// CHECK: [[TMP2:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
98239823
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[B:%.*]] to <8 x i8>
98249824
// CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
9825-
// CHECK: [[VQDMLAL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqdmull.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]]) #8
9826-
// CHECK: [[VQDMLAL_V3_I:%.*]] = call <4 x i32> @llvm.sadd.sat.v4i32(<4 x i32> [[A]], <4 x i32> [[VQDMLAL2_I]]) #8
9825+
// CHECK: [[VQDMLAL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqdmull.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]])
9826+
// CHECK: [[VQDMLAL_V3_I:%.*]] = call <4 x i32> @llvm.sadd.sat.v4i32(<4 x i32> [[A]], <4 x i32> [[VQDMLAL2_I]])
98279827
// CHECK: ret <4 x i32> [[VQDMLAL_V3_I]]
98289828
int32x4_t test_vqdmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t c) {
98299829
return vqdmlal_lane_s16(a, b, c, 3);
@@ -9836,8 +9836,8 @@ int32x4_t test_vqdmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t c) {
98369836
// CHECK: [[TMP2:%.*]] = bitcast <2 x i64> [[A:%.*]] to <16 x i8>
98379837
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[B:%.*]] to <8 x i8>
98389838
// CHECK: [[TMP4:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
9839-
// CHECK: [[VQDMLAL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vqdmull.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]]) #8
9840-
// CHECK: [[VQDMLAL_V3_I:%.*]] = call <2 x i64> @llvm.sadd.sat.v2i64(<2 x i64> [[A]], <2 x i64> [[VQDMLAL2_I]]) #8
9839+
// CHECK: [[VQDMLAL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vqdmull.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]])
9840+
// CHECK: [[VQDMLAL_V3_I:%.*]] = call <2 x i64> @llvm.sadd.sat.v2i64(<2 x i64> [[A]], <2 x i64> [[VQDMLAL2_I]])
98419841
// CHECK: ret <2 x i64> [[VQDMLAL_V3_I]]
98429842
int64x2_t test_vqdmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t c) {
98439843
return vqdmlal_lane_s32(a, b, c, 1);
@@ -9900,8 +9900,8 @@ int64x2_t test_vqdmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c) {
99009900
// CHECK: [[TMP2:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
99019901
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[B:%.*]] to <8 x i8>
99029902
// CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
9903-
// CHECK: [[VQDMLAL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqdmull.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]]) #8
9904-
// CHECK: [[VQDMLSL_V3_I:%.*]] = call <4 x i32> @llvm.ssub.sat.v4i32(<4 x i32> [[A]], <4 x i32> [[VQDMLAL2_I]]) #8
9903+
// CHECK: [[VQDMLAL2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqdmull.v4i32(<4 x i16> [[B]], <4 x i16> [[LANE]])
9904+
// CHECK: [[VQDMLSL_V3_I:%.*]] = call <4 x i32> @llvm.ssub.sat.v4i32(<4 x i32> [[A]], <4 x i32> [[VQDMLAL2_I]])
99059905
// CHECK: ret <4 x i32> [[VQDMLSL_V3_I]]
99069906
int32x4_t test_vqdmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t c) {
99079907
return vqdmlsl_lane_s16(a, b, c, 3);
@@ -9914,8 +9914,8 @@ int32x4_t test_vqdmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t c) {
99149914
// CHECK: [[TMP2:%.*]] = bitcast <2 x i64> [[A:%.*]] to <16 x i8>
99159915
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[B:%.*]] to <8 x i8>
99169916
// CHECK: [[TMP4:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
9917-
// CHECK: [[VQDMLAL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vqdmull.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]]) #8
9918-
// CHECK: [[VQDMLSL_V3_I:%.*]] = call <2 x i64> @llvm.ssub.sat.v2i64(<2 x i64> [[A]], <2 x i64> [[VQDMLAL2_I]]) #8
9917+
// CHECK: [[VQDMLAL2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vqdmull.v2i64(<2 x i32> [[B]], <2 x i32> [[LANE]])
9918+
// CHECK: [[VQDMLSL_V3_I:%.*]] = call <2 x i64> @llvm.ssub.sat.v2i64(<2 x i64> [[A]], <2 x i64> [[VQDMLAL2_I]])
99199919
// CHECK: ret <2 x i64> [[VQDMLSL_V3_I]]
99209920
int64x2_t test_vqdmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t c) {
99219921
return vqdmlsl_lane_s32(a, b, c, 1);
@@ -9995,7 +9995,7 @@ int32x4_t test_vqdmulhq_s32(int32x4_t a, int32x4_t b) {
99959995
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
99969996
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[A:%.*]] to <8 x i8>
99979997
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
9998-
// CHECK: [[VQDMULH_V2_I:%.*]] = call <4 x i16> @llvm.arm.neon.vqdmulh.v4i16(<4 x i16> [[A]], <4 x i16> [[LANE]]) #8
9998+
// CHECK: [[VQDMULH_V2_I:%.*]] = call <4 x i16> @llvm.arm.neon.vqdmulh.v4i16(<4 x i16> [[A]], <4 x i16> [[LANE]])
99999999
// CHECK: [[VQDMULH_V3_I:%.*]] = bitcast <4 x i16> [[VQDMULH_V2_I]] to <8 x i8>
1000010000
// CHECK: ret <4 x i16> [[VQDMULH_V2_I]]
1000110001
int16x4_t test_vqdmulh_lane_s16(int16x4_t a, int16x4_t b) {
@@ -10008,7 +10008,7 @@ int16x4_t test_vqdmulh_lane_s16(int16x4_t a, int16x4_t b) {
1000810008
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
1000910009
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[A:%.*]] to <8 x i8>
1001010010
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
10011-
// CHECK: [[VQDMULH_V2_I:%.*]] = call <2 x i32> @llvm.arm.neon.vqdmulh.v2i32(<2 x i32> [[A]], <2 x i32> [[LANE]]) #8
10011+
// CHECK: [[VQDMULH_V2_I:%.*]] = call <2 x i32> @llvm.arm.neon.vqdmulh.v2i32(<2 x i32> [[A]], <2 x i32> [[LANE]])
1001210012
// CHECK: [[VQDMULH_V3_I:%.*]] = bitcast <2 x i32> [[VQDMULH_V2_I]] to <8 x i8>
1001310013
// CHECK: ret <2 x i32> [[VQDMULH_V2_I]]
1001410014
int32x2_t test_vqdmulh_lane_s32(int32x2_t a, int32x2_t b) {
@@ -10021,7 +10021,7 @@ int32x2_t test_vqdmulh_lane_s32(int32x2_t a, int32x2_t b) {
1002110021
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <8 x i32> <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
1002210022
// CHECK: [[TMP2:%.*]] = bitcast <8 x i16> [[A:%.*]] to <16 x i8>
1002310023
// CHECK: [[TMP3:%.*]] = bitcast <8 x i16> [[LANE]] to <16 x i8>
10024-
// CHECK: [[VQDMULHQ_V2_I:%.*]] = call <8 x i16> @llvm.arm.neon.vqdmulh.v8i16(<8 x i16> [[A]], <8 x i16> [[LANE]]) #8
10024+
// CHECK: [[VQDMULHQ_V2_I:%.*]] = call <8 x i16> @llvm.arm.neon.vqdmulh.v8i16(<8 x i16> [[A]], <8 x i16> [[LANE]])
1002510025
// CHECK: [[VQDMULHQ_V3_I:%.*]] = bitcast <8 x i16> [[VQDMULHQ_V2_I]] to <16 x i8>
1002610026
// CHECK: ret <8 x i16> [[VQDMULHQ_V2_I]]
1002710027
int16x8_t test_vqdmulhq_lane_s16(int16x8_t a, int16x4_t b) {
@@ -10034,7 +10034,7 @@ int16x8_t test_vqdmulhq_lane_s16(int16x8_t a, int16x4_t b) {
1003410034
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <4 x i32> <i32 1, i32 1, i32 1, i32 1>
1003510035
// CHECK: [[TMP2:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
1003610036
// CHECK: [[TMP3:%.*]] = bitcast <4 x i32> [[LANE]] to <16 x i8>
10037-
// CHECK: [[VQDMULHQ_V2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqdmulh.v4i32(<4 x i32> [[A]], <4 x i32> [[LANE]]) #8
10037+
// CHECK: [[VQDMULHQ_V2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqdmulh.v4i32(<4 x i32> [[A]], <4 x i32> [[LANE]])
1003810038
// CHECK: [[VQDMULHQ_V3_I:%.*]] = bitcast <4 x i32> [[VQDMULHQ_V2_I]] to <16 x i8>
1003910039
// CHECK: ret <4 x i32> [[VQDMULHQ_V2_I]]
1004010040
int32x4_t test_vqdmulhq_lane_s32(int32x4_t a, int32x2_t b) {
@@ -10125,7 +10125,7 @@ int64x2_t test_vqdmull_s32(int32x2_t a, int32x2_t b) {
1012510125
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
1012610126
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[A:%.*]] to <8 x i8>
1012710127
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
10128-
// CHECK: [[VQDMULL_V2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqdmull.v4i32(<4 x i16> [[A]], <4 x i16> [[LANE]]) #8
10128+
// CHECK: [[VQDMULL_V2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqdmull.v4i32(<4 x i16> [[A]], <4 x i16> [[LANE]])
1012910129
// CHECK: [[VQDMULL_V3_I:%.*]] = bitcast <4 x i32> [[VQDMULL_V2_I]] to <16 x i8>
1013010130
// CHECK: ret <4 x i32> [[VQDMULL_V2_I]]
1013110131
int32x4_t test_vqdmull_lane_s16(int16x4_t a, int16x4_t b) {
@@ -10138,7 +10138,7 @@ int32x4_t test_vqdmull_lane_s16(int16x4_t a, int16x4_t b) {
1013810138
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
1013910139
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[A:%.*]] to <8 x i8>
1014010140
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
10141-
// CHECK: [[VQDMULL_V2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vqdmull.v2i64(<2 x i32> [[A]], <2 x i32> [[LANE]]) #8
10141+
// CHECK: [[VQDMULL_V2_I:%.*]] = call <2 x i64> @llvm.arm.neon.vqdmull.v2i64(<2 x i32> [[A]], <2 x i32> [[LANE]])
1014210142
// CHECK: [[VQDMULL_V3_I:%.*]] = bitcast <2 x i64> [[VQDMULL_V2_I]] to <16 x i8>
1014310143
// CHECK: ret <2 x i64> [[VQDMULL_V2_I]]
1014410144
int64x2_t test_vqdmull_lane_s32(int32x2_t a, int32x2_t b) {
@@ -10345,7 +10345,7 @@ int32x4_t test_vqrdmulhq_s32(int32x4_t a, int32x4_t b) {
1034510345
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
1034610346
// CHECK: [[TMP2:%.*]] = bitcast <4 x i16> [[A:%.*]] to <8 x i8>
1034710347
// CHECK: [[TMP3:%.*]] = bitcast <4 x i16> [[LANE]] to <8 x i8>
10348-
// CHECK: [[VQRDMULH_V2_I:%.*]] = call <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16> [[A]], <4 x i16> [[LANE]]) #8
10348+
// CHECK: [[VQRDMULH_V2_I:%.*]] = call <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16> [[A]], <4 x i16> [[LANE]])
1034910349
// CHECK: [[VQRDMULH_V3_I:%.*]] = bitcast <4 x i16> [[VQRDMULH_V2_I]] to <8 x i8>
1035010350
// CHECK: ret <4 x i16> [[VQRDMULH_V2_I]]
1035110351
int16x4_t test_vqrdmulh_lane_s16(int16x4_t a, int16x4_t b) {
@@ -10358,7 +10358,7 @@ int16x4_t test_vqrdmulh_lane_s16(int16x4_t a, int16x4_t b) {
1035810358
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> <i32 1, i32 1>
1035910359
// CHECK: [[TMP2:%.*]] = bitcast <2 x i32> [[A:%.*]] to <8 x i8>
1036010360
// CHECK: [[TMP3:%.*]] = bitcast <2 x i32> [[LANE]] to <8 x i8>
10361-
// CHECK: [[VQRDMULH_V2_I:%.*]] = call <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32> [[A]], <2 x i32> [[LANE]]) #8
10361+
// CHECK: [[VQRDMULH_V2_I:%.*]] = call <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32> [[A]], <2 x i32> [[LANE]])
1036210362
// CHECK: [[VQRDMULH_V3_I:%.*]] = bitcast <2 x i32> [[VQRDMULH_V2_I]] to <8 x i8>
1036310363
// CHECK: ret <2 x i32> [[VQRDMULH_V2_I]]
1036410364
int32x2_t test_vqrdmulh_lane_s32(int32x2_t a, int32x2_t b) {
@@ -10371,7 +10371,7 @@ int32x2_t test_vqrdmulh_lane_s32(int32x2_t a, int32x2_t b) {
1037110371
// CHECK: [[LANE:%.*]] = shufflevector <4 x i16> [[TMP1]], <4 x i16> [[TMP1]], <8 x i32> <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
1037210372
// CHECK: [[TMP2:%.*]] = bitcast <8 x i16> [[A:%.*]] to <16 x i8>
1037310373
// CHECK: [[TMP3:%.*]] = bitcast <8 x i16> [[LANE]] to <16 x i8>
10374-
// CHECK: [[VQRDMULHQ_V2_I:%.*]] = call <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16> [[A]], <8 x i16> [[LANE]]) #8
10374+
// CHECK: [[VQRDMULHQ_V2_I:%.*]] = call <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16> [[A]], <8 x i16> [[LANE]])
1037510375
// CHECK: [[VQRDMULHQ_V3_I:%.*]] = bitcast <8 x i16> [[VQRDMULHQ_V2_I]] to <16 x i8>
1037610376
// CHECK: ret <8 x i16> [[VQRDMULHQ_V2_I]]
1037710377
int16x8_t test_vqrdmulhq_lane_s16(int16x8_t a, int16x4_t b) {
@@ -10384,7 +10384,7 @@ int16x8_t test_vqrdmulhq_lane_s16(int16x8_t a, int16x4_t b) {
1038410384
// CHECK: [[LANE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <4 x i32> <i32 1, i32 1, i32 1, i32 1>
1038510385
// CHECK: [[TMP2:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
1038610386
// CHECK: [[TMP3:%.*]] = bitcast <4 x i32> [[LANE]] to <16 x i8>
10387-
// CHECK: [[VQRDMULHQ_V2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32> [[A]], <4 x i32> [[LANE]]) #8
10387+
// CHECK: [[VQRDMULHQ_V2_I:%.*]] = call <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32> [[A]], <4 x i32> [[LANE]])
1038810388
// CHECK: [[VQRDMULHQ_V3_I:%.*]] = bitcast <4 x i32> [[VQRDMULHQ_V2_I]] to <16 x i8>
1038910389
// CHECK: ret <4 x i32> [[VQRDMULHQ_V2_I]]
1039010390
int32x4_t test_vqrdmulhq_lane_s32(int32x4_t a, int32x2_t b) {

0 commit comments

Comments
 (0)