diff options
Diffstat (limited to 'clang/test/CodeGen/aarch64-neon-intrinsics.c')
| -rw-r--r-- | clang/test/CodeGen/aarch64-neon-intrinsics.c | 648 |
1 files changed, 324 insertions, 324 deletions
diff --git a/clang/test/CodeGen/aarch64-neon-intrinsics.c b/clang/test/CodeGen/aarch64-neon-intrinsics.c index cbc2e72fcba..fb9f776c4ad 100644 --- a/clang/test/CodeGen/aarch64-neon-intrinsics.c +++ b/clang/test/CodeGen/aarch64-neon-intrinsics.c @@ -9203,7 +9203,7 @@ poly16x4_t test_vld1_p16(poly16_t const *a) { // CHECK: store { <16 x i8>, <16 x i8> } [[VLD2]], { <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x16x2_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.uint8x16x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 32, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.uint8x16x2_t, %struct.uint8x16x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint8x16x2_t [[TMP5]] uint8x16x2_t test_vld2q_u8(uint8_t const *a) { @@ -9221,7 +9221,7 @@ uint8x16x2_t test_vld2q_u8(uint8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16> } [[VLD2]], { <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x8x2_t, %struct.uint16x8x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint16x8x2_t [[TMP6]] uint16x8x2_t test_vld2q_u16(uint16_t const *a) { @@ -9239,7 +9239,7 @@ uint16x8x2_t test_vld2q_u16(uint16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32> } [[VLD2]], { <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x4x2_t, %struct.uint32x4x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint32x4x2_t [[TMP6]] uint32x4x2_t test_vld2q_u32(uint32_t const *a) { @@ -9257,7 +9257,7 @@ uint32x4x2_t test_vld2q_u32(uint32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64> } [[VLD2]], { <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x2x2_t, %struct.uint64x2x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint64x2x2_t [[TMP6]] uint64x2x2_t test_vld2q_u64(uint64_t const *a) { @@ -9274,7 +9274,7 @@ uint64x2x2_t test_vld2q_u64(uint64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8> } [[VLD2]], { <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x16x2_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.int8x16x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 32, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.int8x16x2_t, %struct.int8x16x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.int8x16x2_t [[TMP5]] int8x16x2_t test_vld2q_s8(int8_t const *a) { @@ -9292,7 +9292,7 @@ int8x16x2_t test_vld2q_s8(int8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16> } [[VLD2]], { <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x8x2_t, %struct.int16x8x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.int16x8x2_t [[TMP6]] int16x8x2_t test_vld2q_s16(int16_t const *a) { @@ -9310,7 +9310,7 @@ int16x8x2_t test_vld2q_s16(int16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32> } [[VLD2]], { <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x4x2_t, %struct.int32x4x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.int32x4x2_t [[TMP6]] int32x4x2_t test_vld2q_s32(int32_t const *a) { @@ -9328,7 +9328,7 @@ int32x4x2_t test_vld2q_s32(int32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64> } [[VLD2]], { <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x2x2_t, %struct.int64x2x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.int64x2x2_t [[TMP6]] int64x2x2_t test_vld2q_s64(int64_t const *a) { @@ -9346,7 +9346,7 @@ int64x2x2_t test_vld2q_s64(int64_t const *a) { // CHECK: store { <8 x half>, <8 x half> } [[VLD2]], { <8 x half>, <8 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x8x2_t, %struct.float16x8x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.float16x8x2_t [[TMP6]] float16x8x2_t test_vld2q_f16(float16_t const *a) { @@ -9364,7 +9364,7 @@ float16x8x2_t test_vld2q_f16(float16_t const *a) { // CHECK: store { <4 x float>, <4 x float> } [[VLD2]], { <4 x float>, <4 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x4x2_t, %struct.float32x4x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.float32x4x2_t [[TMP6]] float32x4x2_t test_vld2q_f32(float32_t const *a) { @@ -9382,7 +9382,7 @@ float32x4x2_t test_vld2q_f32(float32_t const *a) { // CHECK: store { <2 x double>, <2 x double> } [[VLD2]], { <2 x double>, <2 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x2x2_t, %struct.float64x2x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.float64x2x2_t [[TMP6]] float64x2x2_t test_vld2q_f64(float64_t const *a) { @@ -9399,7 +9399,7 @@ float64x2x2_t test_vld2q_f64(float64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8> } [[VLD2]], { <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x16x2_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.poly8x16x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 32, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.poly8x16x2_t, %struct.poly8x16x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly8x16x2_t [[TMP5]] poly8x16x2_t test_vld2q_p8(poly8_t const *a) { @@ -9417,7 +9417,7 @@ poly8x16x2_t test_vld2q_p8(poly8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16> } [[VLD2]], { <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x8x2_t, %struct.poly16x8x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly16x8x2_t [[TMP6]] poly16x8x2_t test_vld2q_p16(poly16_t const *a) { @@ -9434,7 +9434,7 @@ poly16x8x2_t test_vld2q_p16(poly16_t const *a) { // CHECK: store { <8 x i8>, <8 x i8> } [[VLD2]], { <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.uint8x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 16, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.uint8x8x2_t, %struct.uint8x8x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint8x8x2_t [[TMP5]] uint8x8x2_t test_vld2_u8(uint8_t const *a) { @@ -9452,7 +9452,7 @@ uint8x8x2_t test_vld2_u8(uint8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16> } [[VLD2]], { <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x4x2_t, %struct.uint16x4x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint16x4x2_t [[TMP6]] uint16x4x2_t test_vld2_u16(uint16_t const *a) { @@ -9470,7 +9470,7 @@ uint16x4x2_t test_vld2_u16(uint16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32> } [[VLD2]], { <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x2x2_t, %struct.uint32x2x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint32x2x2_t [[TMP6]] uint32x2x2_t test_vld2_u32(uint32_t const *a) { @@ -9488,7 +9488,7 @@ uint32x2x2_t test_vld2_u32(uint32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64> } [[VLD2]], { <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x1x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x1x2_t, %struct.uint64x1x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint64x1x2_t [[TMP6]] uint64x1x2_t test_vld2_u64(uint64_t const *a) { @@ -9505,7 +9505,7 @@ uint64x1x2_t test_vld2_u64(uint64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8> } [[VLD2]], { <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.int8x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 16, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.int8x8x2_t, %struct.int8x8x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.int8x8x2_t [[TMP5]] int8x8x2_t test_vld2_s8(int8_t const *a) { @@ -9523,7 +9523,7 @@ int8x8x2_t test_vld2_s8(int8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16> } [[VLD2]], { <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x4x2_t, %struct.int16x4x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.int16x4x2_t [[TMP6]] int16x4x2_t test_vld2_s16(int16_t const *a) { @@ -9541,7 +9541,7 @@ int16x4x2_t test_vld2_s16(int16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32> } [[VLD2]], { <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x2x2_t, %struct.int32x2x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.int32x2x2_t [[TMP6]] int32x2x2_t test_vld2_s32(int32_t const *a) { @@ -9559,7 +9559,7 @@ int32x2x2_t test_vld2_s32(int32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64> } [[VLD2]], { <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x1x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x1x2_t, %struct.int64x1x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.int64x1x2_t [[TMP6]] int64x1x2_t test_vld2_s64(int64_t const *a) { @@ -9577,7 +9577,7 @@ int64x1x2_t test_vld2_s64(int64_t const *a) { // CHECK: store { <4 x half>, <4 x half> } [[VLD2]], { <4 x half>, <4 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x4x2_t, %struct.float16x4x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.float16x4x2_t [[TMP6]] float16x4x2_t test_vld2_f16(float16_t const *a) { @@ -9595,7 +9595,7 @@ float16x4x2_t test_vld2_f16(float16_t const *a) { // CHECK: store { <2 x float>, <2 x float> } [[VLD2]], { <2 x float>, <2 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x2x2_t, %struct.float32x2x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.float32x2x2_t [[TMP6]] float32x2x2_t test_vld2_f32(float32_t const *a) { @@ -9613,7 +9613,7 @@ float32x2x2_t test_vld2_f32(float32_t const *a) { // CHECK: store { <1 x double>, <1 x double> } [[VLD2]], { <1 x double>, <1 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x1x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x1x2_t, %struct.float64x1x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.float64x1x2_t [[TMP6]] float64x1x2_t test_vld2_f64(float64_t const *a) { @@ -9630,7 +9630,7 @@ float64x1x2_t test_vld2_f64(float64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8> } [[VLD2]], { <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.poly8x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 16, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.poly8x8x2_t, %struct.poly8x8x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly8x8x2_t [[TMP5]] poly8x8x2_t test_vld2_p8(poly8_t const *a) { @@ -9648,7 +9648,7 @@ poly8x8x2_t test_vld2_p8(poly8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16> } [[VLD2]], { <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x4x2_t, %struct.poly16x4x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly16x4x2_t [[TMP6]] poly16x4x2_t test_vld2_p16(poly16_t const *a) { @@ -9665,7 +9665,7 @@ poly16x4x2_t test_vld2_p16(poly16_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8> } [[VLD3]], { <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x16x3_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.uint8x16x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 48, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.uint8x16x3_t, %struct.uint8x16x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint8x16x3_t [[TMP5]] uint8x16x3_t test_vld3q_u8(uint8_t const *a) { @@ -9683,7 +9683,7 @@ uint8x16x3_t test_vld3q_u8(uint8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD3]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x8x3_t, %struct.uint16x8x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint16x8x3_t [[TMP6]] uint16x8x3_t test_vld3q_u16(uint16_t const *a) { @@ -9701,7 +9701,7 @@ uint16x8x3_t test_vld3q_u16(uint16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32>, <4 x i32> } [[VLD3]], { <4 x i32>, <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x4x3_t, %struct.uint32x4x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint32x4x3_t [[TMP6]] uint32x4x3_t test_vld3q_u32(uint32_t const *a) { @@ -9719,7 +9719,7 @@ uint32x4x3_t test_vld3q_u32(uint32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64> } [[VLD3]], { <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x2x3_t, %struct.uint64x2x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint64x2x3_t [[TMP6]] uint64x2x3_t test_vld3q_u64(uint64_t const *a) { @@ -9736,7 +9736,7 @@ uint64x2x3_t test_vld3q_u64(uint64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8> } [[VLD3]], { <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x16x3_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.int8x16x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 48, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.int8x16x3_t, %struct.int8x16x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.int8x16x3_t [[TMP5]] int8x16x3_t test_vld3q_s8(int8_t const *a) { @@ -9754,7 +9754,7 @@ int8x16x3_t test_vld3q_s8(int8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD3]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x8x3_t, %struct.int16x8x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.int16x8x3_t [[TMP6]] int16x8x3_t test_vld3q_s16(int16_t const *a) { @@ -9772,7 +9772,7 @@ int16x8x3_t test_vld3q_s16(int16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32>, <4 x i32> } [[VLD3]], { <4 x i32>, <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x4x3_t, %struct.int32x4x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.int32x4x3_t [[TMP6]] int32x4x3_t test_vld3q_s32(int32_t const *a) { @@ -9790,7 +9790,7 @@ int32x4x3_t test_vld3q_s32(int32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64> } [[VLD3]], { <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x2x3_t, %struct.int64x2x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.int64x2x3_t [[TMP6]] int64x2x3_t test_vld3q_s64(int64_t const *a) { @@ -9808,7 +9808,7 @@ int64x2x3_t test_vld3q_s64(int64_t const *a) { // CHECK: store { <8 x half>, <8 x half>, <8 x half> } [[VLD3]], { <8 x half>, <8 x half>, <8 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x8x3_t, %struct.float16x8x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.float16x8x3_t [[TMP6]] float16x8x3_t test_vld3q_f16(float16_t const *a) { @@ -9826,7 +9826,7 @@ float16x8x3_t test_vld3q_f16(float16_t const *a) { // CHECK: store { <4 x float>, <4 x float>, <4 x float> } [[VLD3]], { <4 x float>, <4 x float>, <4 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x4x3_t, %struct.float32x4x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.float32x4x3_t [[TMP6]] float32x4x3_t test_vld3q_f32(float32_t const *a) { @@ -9844,7 +9844,7 @@ float32x4x3_t test_vld3q_f32(float32_t const *a) { // CHECK: store { <2 x double>, <2 x double>, <2 x double> } [[VLD3]], { <2 x double>, <2 x double>, <2 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x2x3_t, %struct.float64x2x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.float64x2x3_t [[TMP6]] float64x2x3_t test_vld3q_f64(float64_t const *a) { @@ -9861,7 +9861,7 @@ float64x2x3_t test_vld3q_f64(float64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8> } [[VLD3]], { <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x16x3_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.poly8x16x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 48, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.poly8x16x3_t, %struct.poly8x16x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly8x16x3_t [[TMP5]] poly8x16x3_t test_vld3q_p8(poly8_t const *a) { @@ -9879,7 +9879,7 @@ poly8x16x3_t test_vld3q_p8(poly8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD3]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x8x3_t, %struct.poly16x8x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly16x8x3_t [[TMP6]] poly16x8x3_t test_vld3q_p16(poly16_t const *a) { @@ -9896,7 +9896,7 @@ poly16x8x3_t test_vld3q_p16(poly16_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8> } [[VLD3]], { <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.uint8x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 24, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.uint8x8x3_t, %struct.uint8x8x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint8x8x3_t [[TMP5]] uint8x8x3_t test_vld3_u8(uint8_t const *a) { @@ -9914,7 +9914,7 @@ uint8x8x3_t test_vld3_u8(uint8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD3]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x4x3_t, %struct.uint16x4x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint16x4x3_t [[TMP6]] uint16x4x3_t test_vld3_u16(uint16_t const *a) { @@ -9932,7 +9932,7 @@ uint16x4x3_t test_vld3_u16(uint16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32>, <2 x i32> } [[VLD3]], { <2 x i32>, <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x2x3_t, %struct.uint32x2x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint32x2x3_t [[TMP6]] uint32x2x3_t test_vld3_u32(uint32_t const *a) { @@ -9950,7 +9950,7 @@ uint32x2x3_t test_vld3_u32(uint32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64> } [[VLD3]], { <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x1x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x1x3_t, %struct.uint64x1x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint64x1x3_t [[TMP6]] uint64x1x3_t test_vld3_u64(uint64_t const *a) { @@ -9967,7 +9967,7 @@ uint64x1x3_t test_vld3_u64(uint64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8> } [[VLD3]], { <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.int8x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 24, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.int8x8x3_t, %struct.int8x8x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.int8x8x3_t [[TMP5]] int8x8x3_t test_vld3_s8(int8_t const *a) { @@ -9985,7 +9985,7 @@ int8x8x3_t test_vld3_s8(int8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD3]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x4x3_t, %struct.int16x4x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.int16x4x3_t [[TMP6]] int16x4x3_t test_vld3_s16(int16_t const *a) { @@ -10003,7 +10003,7 @@ int16x4x3_t test_vld3_s16(int16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32>, <2 x i32> } [[VLD3]], { <2 x i32>, <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x2x3_t, %struct.int32x2x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.int32x2x3_t [[TMP6]] int32x2x3_t test_vld3_s32(int32_t const *a) { @@ -10021,7 +10021,7 @@ int32x2x3_t test_vld3_s32(int32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64> } [[VLD3]], { <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x1x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x1x3_t, %struct.int64x1x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.int64x1x3_t [[TMP6]] int64x1x3_t test_vld3_s64(int64_t const *a) { @@ -10039,7 +10039,7 @@ int64x1x3_t test_vld3_s64(int64_t const *a) { // CHECK: store { <4 x half>, <4 x half>, <4 x half> } [[VLD3]], { <4 x half>, <4 x half>, <4 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x4x3_t, %struct.float16x4x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.float16x4x3_t [[TMP6]] float16x4x3_t test_vld3_f16(float16_t const *a) { @@ -10057,7 +10057,7 @@ float16x4x3_t test_vld3_f16(float16_t const *a) { // CHECK: store { <2 x float>, <2 x float>, <2 x float> } [[VLD3]], { <2 x float>, <2 x float>, <2 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x2x3_t, %struct.float32x2x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.float32x2x3_t [[TMP6]] float32x2x3_t test_vld3_f32(float32_t const *a) { @@ -10075,7 +10075,7 @@ float32x2x3_t test_vld3_f32(float32_t const *a) { // CHECK: store { <1 x double>, <1 x double>, <1 x double> } [[VLD3]], { <1 x double>, <1 x double>, <1 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x1x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x1x3_t, %struct.float64x1x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.float64x1x3_t [[TMP6]] float64x1x3_t test_vld3_f64(float64_t const *a) { @@ -10092,7 +10092,7 @@ float64x1x3_t test_vld3_f64(float64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8> } [[VLD3]], { <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.poly8x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 24, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.poly8x8x3_t, %struct.poly8x8x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly8x8x3_t [[TMP5]] poly8x8x3_t test_vld3_p8(poly8_t const *a) { @@ -10110,7 +10110,7 @@ poly8x8x3_t test_vld3_p8(poly8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD3]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x4x3_t, %struct.poly16x4x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly16x4x3_t [[TMP6]] poly16x4x3_t test_vld3_p16(poly16_t const *a) { @@ -10127,7 +10127,7 @@ poly16x4x3_t test_vld3_p16(poly16_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } [[VLD4]], { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x16x4_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.uint8x16x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 64, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.uint8x16x4_t, %struct.uint8x16x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint8x16x4_t [[TMP5]] uint8x16x4_t test_vld4q_u8(uint8_t const *a) { @@ -10145,7 +10145,7 @@ uint8x16x4_t test_vld4q_u8(uint8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD4]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x8x4_t, %struct.uint16x8x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint16x8x4_t [[TMP6]] uint16x8x4_t test_vld4q_u16(uint16_t const *a) { @@ -10163,7 +10163,7 @@ uint16x8x4_t test_vld4q_u16(uint16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> } [[VLD4]], { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x4x4_t, %struct.uint32x4x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint32x4x4_t [[TMP6]] uint32x4x4_t test_vld4q_u32(uint32_t const *a) { @@ -10181,7 +10181,7 @@ uint32x4x4_t test_vld4q_u32(uint32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> } [[VLD4]], { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x2x4_t, %struct.uint64x2x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint64x2x4_t [[TMP6]] uint64x2x4_t test_vld4q_u64(uint64_t const *a) { @@ -10198,7 +10198,7 @@ uint64x2x4_t test_vld4q_u64(uint64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } [[VLD4]], { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x16x4_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.int8x16x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 64, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.int8x16x4_t, %struct.int8x16x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.int8x16x4_t [[TMP5]] int8x16x4_t test_vld4q_s8(int8_t const *a) { @@ -10216,7 +10216,7 @@ int8x16x4_t test_vld4q_s8(int8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD4]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x8x4_t, %struct.int16x8x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.int16x8x4_t [[TMP6]] int16x8x4_t test_vld4q_s16(int16_t const *a) { @@ -10234,7 +10234,7 @@ int16x8x4_t test_vld4q_s16(int16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> } [[VLD4]], { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x4x4_t, %struct.int32x4x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.int32x4x4_t [[TMP6]] int32x4x4_t test_vld4q_s32(int32_t const *a) { @@ -10252,7 +10252,7 @@ int32x4x4_t test_vld4q_s32(int32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> } [[VLD4]], { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x2x4_t, %struct.int64x2x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.int64x2x4_t [[TMP6]] int64x2x4_t test_vld4q_s64(int64_t const *a) { @@ -10270,7 +10270,7 @@ int64x2x4_t test_vld4q_s64(int64_t const *a) { // CHECK: store { <8 x half>, <8 x half>, <8 x half>, <8 x half> } [[VLD4]], { <8 x half>, <8 x half>, <8 x half>, <8 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x8x4_t, %struct.float16x8x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.float16x8x4_t [[TMP6]] float16x8x4_t test_vld4q_f16(float16_t const *a) { @@ -10288,7 +10288,7 @@ float16x8x4_t test_vld4q_f16(float16_t const *a) { // CHECK: store { <4 x float>, <4 x float>, <4 x float>, <4 x float> } [[VLD4]], { <4 x float>, <4 x float>, <4 x float>, <4 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x4x4_t, %struct.float32x4x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.float32x4x4_t [[TMP6]] float32x4x4_t test_vld4q_f32(float32_t const *a) { @@ -10306,7 +10306,7 @@ float32x4x4_t test_vld4q_f32(float32_t const *a) { // CHECK: store { <2 x double>, <2 x double>, <2 x double>, <2 x double> } [[VLD4]], { <2 x double>, <2 x double>, <2 x double>, <2 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x2x4_t, %struct.float64x2x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.float64x2x4_t [[TMP6]] float64x2x4_t test_vld4q_f64(float64_t const *a) { @@ -10323,7 +10323,7 @@ float64x2x4_t test_vld4q_f64(float64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } [[VLD4]], { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x16x4_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.poly8x16x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP3]], i8* align 16 [[TMP4]], i64 64, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.poly8x16x4_t, %struct.poly8x16x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly8x16x4_t [[TMP5]] poly8x16x4_t test_vld4q_p8(poly8_t const *a) { @@ -10341,7 +10341,7 @@ poly8x16x4_t test_vld4q_p8(poly8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD4]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x8x4_t, %struct.poly16x8x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly16x8x4_t [[TMP6]] poly16x8x4_t test_vld4q_p16(poly16_t const *a) { @@ -10358,7 +10358,7 @@ poly16x8x4_t test_vld4q_p16(poly16_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } [[VLD4]], { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.uint8x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 32, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.uint8x8x4_t, %struct.uint8x8x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint8x8x4_t [[TMP5]] uint8x8x4_t test_vld4_u8(uint8_t const *a) { @@ -10376,7 +10376,7 @@ uint8x8x4_t test_vld4_u8(uint8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD4]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x4x4_t, %struct.uint16x4x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint16x4x4_t [[TMP6]] uint16x4x4_t test_vld4_u16(uint16_t const *a) { @@ -10394,7 +10394,7 @@ uint16x4x4_t test_vld4_u16(uint16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } [[VLD4]], { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x2x4_t, %struct.uint32x2x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint32x2x4_t [[TMP6]] uint32x2x4_t test_vld4_u32(uint32_t const *a) { @@ -10412,7 +10412,7 @@ uint32x2x4_t test_vld4_u32(uint32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> } [[VLD4]], { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x1x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x1x4_t, %struct.uint64x1x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint64x1x4_t [[TMP6]] uint64x1x4_t test_vld4_u64(uint64_t const *a) { @@ -10429,7 +10429,7 @@ uint64x1x4_t test_vld4_u64(uint64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } [[VLD4]], { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.int8x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 32, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.int8x8x4_t, %struct.int8x8x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.int8x8x4_t [[TMP5]] int8x8x4_t test_vld4_s8(int8_t const *a) { @@ -10447,7 +10447,7 @@ int8x8x4_t test_vld4_s8(int8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD4]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x4x4_t, %struct.int16x4x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.int16x4x4_t [[TMP6]] int16x4x4_t test_vld4_s16(int16_t const *a) { @@ -10465,7 +10465,7 @@ int16x4x4_t test_vld4_s16(int16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } [[VLD4]], { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x2x4_t, %struct.int32x2x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.int32x2x4_t [[TMP6]] int32x2x4_t test_vld4_s32(int32_t const *a) { @@ -10483,7 +10483,7 @@ int32x2x4_t test_vld4_s32(int32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> } [[VLD4]], { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x1x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x1x4_t, %struct.int64x1x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.int64x1x4_t [[TMP6]] int64x1x4_t test_vld4_s64(int64_t const *a) { @@ -10501,7 +10501,7 @@ int64x1x4_t test_vld4_s64(int64_t const *a) { // CHECK: store { <4 x half>, <4 x half>, <4 x half>, <4 x half> } [[VLD4]], { <4 x half>, <4 x half>, <4 x half>, <4 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x4x4_t, %struct.float16x4x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.float16x4x4_t [[TMP6]] float16x4x4_t test_vld4_f16(float16_t const *a) { @@ -10519,7 +10519,7 @@ float16x4x4_t test_vld4_f16(float16_t const *a) { // CHECK: store { <2 x float>, <2 x float>, <2 x float>, <2 x float> } [[VLD4]], { <2 x float>, <2 x float>, <2 x float>, <2 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x2x4_t, %struct.float32x2x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.float32x2x4_t [[TMP6]] float32x2x4_t test_vld4_f32(float32_t const *a) { @@ -10537,7 +10537,7 @@ float32x2x4_t test_vld4_f32(float32_t const *a) { // CHECK: store { <1 x double>, <1 x double>, <1 x double>, <1 x double> } [[VLD4]], { <1 x double>, <1 x double>, <1 x double>, <1 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x1x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x1x4_t, %struct.float64x1x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.float64x1x4_t [[TMP6]] float64x1x4_t test_vld4_f64(float64_t const *a) { @@ -10554,7 +10554,7 @@ float64x1x4_t test_vld4_f64(float64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } [[VLD4]], { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP2]] // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP4:%.*]] = bitcast %struct.poly8x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP3]], i8* [[TMP4]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP3]], i8* align 8 [[TMP4]], i64 32, i1 false) // CHECK: [[TMP5:%.*]] = load %struct.poly8x8x4_t, %struct.poly8x8x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly8x8x4_t [[TMP5]] poly8x8x4_t test_vld4_p8(poly8_t const *a) { @@ -10572,7 +10572,7 @@ poly8x8x4_t test_vld4_p8(poly8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD4]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x4x4_t, %struct.poly16x4x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly16x4x4_t [[TMP6]] poly16x4x4_t test_vld4_p16(poly16_t const *a) { @@ -10854,7 +10854,7 @@ void test_vst1_p16(poly16_t *a, poly16x4_t b) { // CHECK: store [2 x <16 x i8>] [[B]].coerce, [2 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x16x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x16x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x16x2_t, %struct.uint8x16x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <16 x i8>], [2 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -10874,7 +10874,7 @@ void test_vst2q_u8(uint8_t *a, uint8x16x2_t b) { // CHECK: store [2 x <8 x i16>] [[B]].coerce, [2 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x8x2_t, %struct.uint16x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i16>], [2 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -10899,7 +10899,7 @@ void test_vst2q_u16(uint16_t *a, uint16x8x2_t b) { // CHECK: store [2 x <4 x i32>] [[B]].coerce, [2 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x4x2_t, %struct.uint32x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i32>], [2 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -10924,7 +10924,7 @@ void test_vst2q_u32(uint32_t *a, uint32x4x2_t b) { // CHECK: store [2 x <2 x i64>] [[B]].coerce, [2 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x2x2_t, %struct.uint64x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i64>], [2 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -10949,7 +10949,7 @@ void test_vst2q_u64(uint64_t *a, uint64x2x2_t b) { // CHECK: store [2 x <16 x i8>] [[B]].coerce, [2 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x16x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x16x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x16x2_t, %struct.int8x16x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <16 x i8>], [2 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -10969,7 +10969,7 @@ void test_vst2q_s8(int8_t *a, int8x16x2_t b) { // CHECK: store [2 x <8 x i16>] [[B]].coerce, [2 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x8x2_t, %struct.int16x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i16>], [2 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -10994,7 +10994,7 @@ void test_vst2q_s16(int16_t *a, int16x8x2_t b) { // CHECK: store [2 x <4 x i32>] [[B]].coerce, [2 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x4x2_t, %struct.int32x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i32>], [2 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -11019,7 +11019,7 @@ void test_vst2q_s32(int32_t *a, int32x4x2_t b) { // CHECK: store [2 x <2 x i64>] [[B]].coerce, [2 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x2x2_t, %struct.int64x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i64>], [2 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -11044,7 +11044,7 @@ void test_vst2q_s64(int64_t *a, int64x2x2_t b) { // CHECK: store [2 x <8 x half>] [[B]].coerce, [2 x <8 x half>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x8x2_t, %struct.float16x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL]], i64 0, i64 0 @@ -11069,7 +11069,7 @@ void test_vst2q_f16(float16_t *a, float16x8x2_t b) { // CHECK: store [2 x <4 x float>] [[B]].coerce, [2 x <4 x float>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x4x2_t, %struct.float32x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x float>], [2 x <4 x float>]* [[VAL]], i64 0, i64 0 @@ -11094,7 +11094,7 @@ void test_vst2q_f32(float32_t *a, float32x4x2_t b) { // CHECK: store [2 x <2 x double>] [[B]].coerce, [2 x <2 x double>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x2x2_t, %struct.float64x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x double>], [2 x <2 x double>]* [[VAL]], i64 0, i64 0 @@ -11119,7 +11119,7 @@ void test_vst2q_f64(float64_t *a, float64x2x2_t b) { // CHECK: store [2 x <16 x i8>] [[B]].coerce, [2 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x16x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x16x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x16x2_t, %struct.poly8x16x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <16 x i8>], [2 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -11139,7 +11139,7 @@ void test_vst2q_p8(poly8_t *a, poly8x16x2_t b) { // CHECK: store [2 x <8 x i16>] [[B]].coerce, [2 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x8x2_t, %struct.poly16x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i16>], [2 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -11164,7 +11164,7 @@ void test_vst2q_p16(poly16_t *a, poly16x8x2_t b) { // CHECK: store [2 x <8 x i8>] [[B]].coerce, [2 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x8x2_t, %struct.uint8x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i8>], [2 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -11184,7 +11184,7 @@ void test_vst2_u8(uint8_t *a, uint8x8x2_t b) { // CHECK: store [2 x <4 x i16>] [[B]].coerce, [2 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x4x2_t, %struct.uint16x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i16>], [2 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -11209,7 +11209,7 @@ void test_vst2_u16(uint16_t *a, uint16x4x2_t b) { // CHECK: store [2 x <2 x i32>] [[B]].coerce, [2 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x2x2_t, %struct.uint32x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i32>], [2 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -11234,7 +11234,7 @@ void test_vst2_u32(uint32_t *a, uint32x2x2_t b) { // CHECK: store [2 x <1 x i64>] [[B]].coerce, [2 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x1x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x1x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x1x2_t, %struct.uint64x1x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <1 x i64>], [2 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -11259,7 +11259,7 @@ void test_vst2_u64(uint64_t *a, uint64x1x2_t b) { // CHECK: store [2 x <8 x i8>] [[B]].coerce, [2 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x8x2_t, %struct.int8x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i8>], [2 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -11279,7 +11279,7 @@ void test_vst2_s8(int8_t *a, int8x8x2_t b) { // CHECK: store [2 x <4 x i16>] [[B]].coerce, [2 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x4x2_t, %struct.int16x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i16>], [2 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -11304,7 +11304,7 @@ void test_vst2_s16(int16_t *a, int16x4x2_t b) { // CHECK: store [2 x <2 x i32>] [[B]].coerce, [2 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x2x2_t, %struct.int32x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i32>], [2 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -11329,7 +11329,7 @@ void test_vst2_s32(int32_t *a, int32x2x2_t b) { // CHECK: store [2 x <1 x i64>] [[B]].coerce, [2 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x1x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x1x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x1x2_t, %struct.int64x1x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <1 x i64>], [2 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -11354,7 +11354,7 @@ void test_vst2_s64(int64_t *a, int64x1x2_t b) { // CHECK: store [2 x <4 x half>] [[B]].coerce, [2 x <4 x half>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x4x2_t, %struct.float16x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL]], i64 0, i64 0 @@ -11379,7 +11379,7 @@ void test_vst2_f16(float16_t *a, float16x4x2_t b) { // CHECK: store [2 x <2 x float>] [[B]].coerce, [2 x <2 x float>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x2x2_t, %struct.float32x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x float>], [2 x <2 x float>]* [[VAL]], i64 0, i64 0 @@ -11404,7 +11404,7 @@ void test_vst2_f32(float32_t *a, float32x2x2_t b) { // CHECK: store [2 x <1 x double>] [[B]].coerce, [2 x <1 x double>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x1x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x1x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x1x2_t, %struct.float64x1x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <1 x double>], [2 x <1 x double>]* [[VAL]], i64 0, i64 0 @@ -11429,7 +11429,7 @@ void test_vst2_f64(float64_t *a, float64x1x2_t b) { // CHECK: store [2 x <8 x i8>] [[B]].coerce, [2 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x8x2_t, %struct.poly8x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i8>], [2 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -11449,7 +11449,7 @@ void test_vst2_p8(poly8_t *a, poly8x8x2_t b) { // CHECK: store [2 x <4 x i16>] [[B]].coerce, [2 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x4x2_t, %struct.poly16x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i16>], [2 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -11474,7 +11474,7 @@ void test_vst2_p16(poly16_t *a, poly16x4x2_t b) { // CHECK: store [3 x <16 x i8>] [[B]].coerce, [3 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x16x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x16x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x16x3_t, %struct.uint8x16x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <16 x i8>], [3 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -11497,7 +11497,7 @@ void test_vst3q_u8(uint8_t *a, uint8x16x3_t b) { // CHECK: store [3 x <8 x i16>] [[B]].coerce, [3 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x8x3_t, %struct.uint16x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i16>], [3 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -11527,7 +11527,7 @@ void test_vst3q_u16(uint16_t *a, uint16x8x3_t b) { // CHECK: store [3 x <4 x i32>] [[B]].coerce, [3 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x4x3_t, %struct.uint32x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i32>], [3 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -11557,7 +11557,7 @@ void test_vst3q_u32(uint32_t *a, uint32x4x3_t b) { // CHECK: store [3 x <2 x i64>] [[B]].coerce, [3 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x2x3_t, %struct.uint64x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i64>], [3 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -11587,7 +11587,7 @@ void test_vst3q_u64(uint64_t *a, uint64x2x3_t b) { // CHECK: store [3 x <16 x i8>] [[B]].coerce, [3 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x16x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x16x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x16x3_t, %struct.int8x16x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <16 x i8>], [3 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -11610,7 +11610,7 @@ void test_vst3q_s8(int8_t *a, int8x16x3_t b) { // CHECK: store [3 x <8 x i16>] [[B]].coerce, [3 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x8x3_t, %struct.int16x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i16>], [3 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -11640,7 +11640,7 @@ void test_vst3q_s16(int16_t *a, int16x8x3_t b) { // CHECK: store [3 x <4 x i32>] [[B]].coerce, [3 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x4x3_t, %struct.int32x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i32>], [3 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -11670,7 +11670,7 @@ void test_vst3q_s32(int32_t *a, int32x4x3_t b) { // CHECK: store [3 x <2 x i64>] [[B]].coerce, [3 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x2x3_t, %struct.int64x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i64>], [3 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -11700,7 +11700,7 @@ void test_vst3q_s64(int64_t *a, int64x2x3_t b) { // CHECK: store [3 x <8 x half>] [[B]].coerce, [3 x <8 x half>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x8x3_t, %struct.float16x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL]], i64 0, i64 0 @@ -11730,7 +11730,7 @@ void test_vst3q_f16(float16_t *a, float16x8x3_t b) { // CHECK: store [3 x <4 x float>] [[B]].coerce, [3 x <4 x float>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x4x3_t, %struct.float32x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x float>], [3 x <4 x float>]* [[VAL]], i64 0, i64 0 @@ -11760,7 +11760,7 @@ void test_vst3q_f32(float32_t *a, float32x4x3_t b) { // CHECK: store [3 x <2 x double>] [[B]].coerce, [3 x <2 x double>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x2x3_t, %struct.float64x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x double>], [3 x <2 x double>]* [[VAL]], i64 0, i64 0 @@ -11790,7 +11790,7 @@ void test_vst3q_f64(float64_t *a, float64x2x3_t b) { // CHECK: store [3 x <16 x i8>] [[B]].coerce, [3 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x16x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x16x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x16x3_t, %struct.poly8x16x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <16 x i8>], [3 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -11813,7 +11813,7 @@ void test_vst3q_p8(poly8_t *a, poly8x16x3_t b) { // CHECK: store [3 x <8 x i16>] [[B]].coerce, [3 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x8x3_t, %struct.poly16x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i16>], [3 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -11843,7 +11843,7 @@ void test_vst3q_p16(poly16_t *a, poly16x8x3_t b) { // CHECK: store [3 x <8 x i8>] [[B]].coerce, [3 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x8x3_t, %struct.uint8x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i8>], [3 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -11866,7 +11866,7 @@ void test_vst3_u8(uint8_t *a, uint8x8x3_t b) { // CHECK: store [3 x <4 x i16>] [[B]].coerce, [3 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x4x3_t, %struct.uint16x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i16>], [3 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -11896,7 +11896,7 @@ void test_vst3_u16(uint16_t *a, uint16x4x3_t b) { // CHECK: store [3 x <2 x i32>] [[B]].coerce, [3 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x2x3_t, %struct.uint32x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i32>], [3 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -11926,7 +11926,7 @@ void test_vst3_u32(uint32_t *a, uint32x2x3_t b) { // CHECK: store [3 x <1 x i64>] [[B]].coerce, [3 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x1x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x1x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x1x3_t, %struct.uint64x1x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <1 x i64>], [3 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -11956,7 +11956,7 @@ void test_vst3_u64(uint64_t *a, uint64x1x3_t b) { // CHECK: store [3 x <8 x i8>] [[B]].coerce, [3 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x8x3_t, %struct.int8x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i8>], [3 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -11979,7 +11979,7 @@ void test_vst3_s8(int8_t *a, int8x8x3_t b) { // CHECK: store [3 x <4 x i16>] [[B]].coerce, [3 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x4x3_t, %struct.int16x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i16>], [3 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -12009,7 +12009,7 @@ void test_vst3_s16(int16_t *a, int16x4x3_t b) { // CHECK: store [3 x <2 x i32>] [[B]].coerce, [3 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x2x3_t, %struct.int32x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i32>], [3 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -12039,7 +12039,7 @@ void test_vst3_s32(int32_t *a, int32x2x3_t b) { // CHECK: store [3 x <1 x i64>] [[B]].coerce, [3 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x1x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x1x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x1x3_t, %struct.int64x1x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <1 x i64>], [3 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -12069,7 +12069,7 @@ void test_vst3_s64(int64_t *a, int64x1x3_t b) { // CHECK: store [3 x <4 x half>] [[B]].coerce, [3 x <4 x half>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x4x3_t, %struct.float16x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL]], i64 0, i64 0 @@ -12099,7 +12099,7 @@ void test_vst3_f16(float16_t *a, float16x4x3_t b) { // CHECK: store [3 x <2 x float>] [[B]].coerce, [3 x <2 x float>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x2x3_t, %struct.float32x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x float>], [3 x <2 x float>]* [[VAL]], i64 0, i64 0 @@ -12129,7 +12129,7 @@ void test_vst3_f32(float32_t *a, float32x2x3_t b) { // CHECK: store [3 x <1 x double>] [[B]].coerce, [3 x <1 x double>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x1x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x1x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x1x3_t, %struct.float64x1x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <1 x double>], [3 x <1 x double>]* [[VAL]], i64 0, i64 0 @@ -12159,7 +12159,7 @@ void test_vst3_f64(float64_t *a, float64x1x3_t b) { // CHECK: store [3 x <8 x i8>] [[B]].coerce, [3 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x8x3_t, %struct.poly8x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i8>], [3 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -12182,7 +12182,7 @@ void test_vst3_p8(poly8_t *a, poly8x8x3_t b) { // CHECK: store [3 x <4 x i16>] [[B]].coerce, [3 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x4x3_t, %struct.poly16x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i16>], [3 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -12212,7 +12212,7 @@ void test_vst3_p16(poly16_t *a, poly16x4x3_t b) { // CHECK: store [4 x <16 x i8>] [[B]].coerce, [4 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x16x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x16x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x16x4_t, %struct.uint8x16x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <16 x i8>], [4 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -12238,7 +12238,7 @@ void test_vst4q_u8(uint8_t *a, uint8x16x4_t b) { // CHECK: store [4 x <8 x i16>] [[B]].coerce, [4 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x8x4_t, %struct.uint16x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i16>], [4 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -12273,7 +12273,7 @@ void test_vst4q_u16(uint16_t *a, uint16x8x4_t b) { // CHECK: store [4 x <4 x i32>] [[B]].coerce, [4 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x4x4_t, %struct.uint32x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i32>], [4 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -12308,7 +12308,7 @@ void test_vst4q_u32(uint32_t *a, uint32x4x4_t b) { // CHECK: store [4 x <2 x i64>] [[B]].coerce, [4 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x2x4_t, %struct.uint64x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i64>], [4 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -12343,7 +12343,7 @@ void test_vst4q_u64(uint64_t *a, uint64x2x4_t b) { // CHECK: store [4 x <16 x i8>] [[B]].coerce, [4 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x16x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x16x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x16x4_t, %struct.int8x16x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <16 x i8>], [4 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -12369,7 +12369,7 @@ void test_vst4q_s8(int8_t *a, int8x16x4_t b) { // CHECK: store [4 x <8 x i16>] [[B]].coerce, [4 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x8x4_t, %struct.int16x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i16>], [4 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -12404,7 +12404,7 @@ void test_vst4q_s16(int16_t *a, int16x8x4_t b) { // CHECK: store [4 x <4 x i32>] [[B]].coerce, [4 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x4x4_t, %struct.int32x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i32>], [4 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -12439,7 +12439,7 @@ void test_vst4q_s32(int32_t *a, int32x4x4_t b) { // CHECK: store [4 x <2 x i64>] [[B]].coerce, [4 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x2x4_t, %struct.int64x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i64>], [4 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -12474,7 +12474,7 @@ void test_vst4q_s64(int64_t *a, int64x2x4_t b) { // CHECK: store [4 x <8 x half>] [[B]].coerce, [4 x <8 x half>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x8x4_t, %struct.float16x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL]], i64 0, i64 0 @@ -12509,7 +12509,7 @@ void test_vst4q_f16(float16_t *a, float16x8x4_t b) { // CHECK: store [4 x <4 x float>] [[B]].coerce, [4 x <4 x float>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x4x4_t, %struct.float32x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x float>], [4 x <4 x float>]* [[VAL]], i64 0, i64 0 @@ -12544,7 +12544,7 @@ void test_vst4q_f32(float32_t *a, float32x4x4_t b) { // CHECK: store [4 x <2 x double>] [[B]].coerce, [4 x <2 x double>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x2x4_t, %struct.float64x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x double>], [4 x <2 x double>]* [[VAL]], i64 0, i64 0 @@ -12579,7 +12579,7 @@ void test_vst4q_f64(float64_t *a, float64x2x4_t b) { // CHECK: store [4 x <16 x i8>] [[B]].coerce, [4 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x16x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x16x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x16x4_t, %struct.poly8x16x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <16 x i8>], [4 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -12605,7 +12605,7 @@ void test_vst4q_p8(poly8_t *a, poly8x16x4_t b) { // CHECK: store [4 x <8 x i16>] [[B]].coerce, [4 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x8x4_t, %struct.poly16x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i16>], [4 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -12640,7 +12640,7 @@ void test_vst4q_p16(poly16_t *a, poly16x8x4_t b) { // CHECK: store [4 x <8 x i8>] [[B]].coerce, [4 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x8x4_t, %struct.uint8x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i8>], [4 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -12666,7 +12666,7 @@ void test_vst4_u8(uint8_t *a, uint8x8x4_t b) { // CHECK: store [4 x <4 x i16>] [[B]].coerce, [4 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x4x4_t, %struct.uint16x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i16>], [4 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -12701,7 +12701,7 @@ void test_vst4_u16(uint16_t *a, uint16x4x4_t b) { // CHECK: store [4 x <2 x i32>] [[B]].coerce, [4 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x2x4_t, %struct.uint32x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i32>], [4 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -12736,7 +12736,7 @@ void test_vst4_u32(uint32_t *a, uint32x2x4_t b) { // CHECK: store [4 x <1 x i64>] [[B]].coerce, [4 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x1x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x1x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x1x4_t, %struct.uint64x1x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <1 x i64>], [4 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -12771,7 +12771,7 @@ void test_vst4_u64(uint64_t *a, uint64x1x4_t b) { // CHECK: store [4 x <8 x i8>] [[B]].coerce, [4 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x8x4_t, %struct.int8x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i8>], [4 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -12797,7 +12797,7 @@ void test_vst4_s8(int8_t *a, int8x8x4_t b) { // CHECK: store [4 x <4 x i16>] [[B]].coerce, [4 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x4x4_t, %struct.int16x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i16>], [4 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -12832,7 +12832,7 @@ void test_vst4_s16(int16_t *a, int16x4x4_t b) { // CHECK: store [4 x <2 x i32>] [[B]].coerce, [4 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x2x4_t, %struct.int32x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i32>], [4 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -12867,7 +12867,7 @@ void test_vst4_s32(int32_t *a, int32x2x4_t b) { // CHECK: store [4 x <1 x i64>] [[B]].coerce, [4 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x1x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x1x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x1x4_t, %struct.int64x1x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <1 x i64>], [4 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -12902,7 +12902,7 @@ void test_vst4_s64(int64_t *a, int64x1x4_t b) { // CHECK: store [4 x <4 x half>] [[B]].coerce, [4 x <4 x half>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x4x4_t, %struct.float16x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL]], i64 0, i64 0 @@ -12937,7 +12937,7 @@ void test_vst4_f16(float16_t *a, float16x4x4_t b) { // CHECK: store [4 x <2 x float>] [[B]].coerce, [4 x <2 x float>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x2x4_t, %struct.float32x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x float>], [4 x <2 x float>]* [[VAL]], i64 0, i64 0 @@ -12972,7 +12972,7 @@ void test_vst4_f32(float32_t *a, float32x2x4_t b) { // CHECK: store [4 x <1 x double>] [[B]].coerce, [4 x <1 x double>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x1x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x1x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x1x4_t, %struct.float64x1x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <1 x double>], [4 x <1 x double>]* [[VAL]], i64 0, i64 0 @@ -13007,7 +13007,7 @@ void test_vst4_f64(float64_t *a, float64x1x4_t b) { // CHECK: store [4 x <8 x i8>] [[B]].coerce, [4 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x8x4_t, %struct.poly8x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i8>], [4 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -13033,7 +13033,7 @@ void test_vst4_p8(poly8_t *a, poly8x8x4_t b) { // CHECK: store [4 x <4 x i16>] [[B]].coerce, [4 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x4x4_t, %struct.poly16x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i16>], [4 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -13070,7 +13070,7 @@ void test_vst4_p16(poly16_t *a, poly16x4x4_t b) { // CHECK: store { <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.uint8x16x2_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x16x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 32, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.uint8x16x2_t, %struct.uint8x16x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint8x16x2_t [[TMP4]] uint8x16x2_t test_vld1q_u8_x2(uint8_t const *a) { @@ -13088,7 +13088,7 @@ uint8x16x2_t test_vld1q_u8_x2(uint8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x8x2_t, %struct.uint16x8x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint16x8x2_t [[TMP6]] uint16x8x2_t test_vld1q_u16_x2(uint16_t const *a) { @@ -13106,7 +13106,7 @@ uint16x8x2_t test_vld1q_u16_x2(uint16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32> } [[VLD1XN]], { <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x4x2_t, %struct.uint32x4x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint32x4x2_t [[TMP6]] uint32x4x2_t test_vld1q_u32_x2(uint32_t const *a) { @@ -13124,7 +13124,7 @@ uint32x4x2_t test_vld1q_u32_x2(uint32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x2x2_t, %struct.uint64x2x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint64x2x2_t [[TMP6]] uint64x2x2_t test_vld1q_u64_x2(uint64_t const *a) { @@ -13140,7 +13140,7 @@ uint64x2x2_t test_vld1q_u64_x2(uint64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.int8x16x2_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x16x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 32, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.int8x16x2_t, %struct.int8x16x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.int8x16x2_t [[TMP4]] int8x16x2_t test_vld1q_s8_x2(int8_t const *a) { @@ -13158,7 +13158,7 @@ int8x16x2_t test_vld1q_s8_x2(int8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x8x2_t, %struct.int16x8x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.int16x8x2_t [[TMP6]] int16x8x2_t test_vld1q_s16_x2(int16_t const *a) { @@ -13176,7 +13176,7 @@ int16x8x2_t test_vld1q_s16_x2(int16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32> } [[VLD1XN]], { <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x4x2_t, %struct.int32x4x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.int32x4x2_t [[TMP6]] int32x4x2_t test_vld1q_s32_x2(int32_t const *a) { @@ -13194,7 +13194,7 @@ int32x4x2_t test_vld1q_s32_x2(int32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x2x2_t, %struct.int64x2x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.int64x2x2_t [[TMP6]] int64x2x2_t test_vld1q_s64_x2(int64_t const *a) { @@ -13212,7 +13212,7 @@ int64x2x2_t test_vld1q_s64_x2(int64_t const *a) { // CHECK: store { <8 x half>, <8 x half> } [[VLD1XN]], { <8 x half>, <8 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x8x2_t, %struct.float16x8x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.float16x8x2_t [[TMP6]] float16x8x2_t test_vld1q_f16_x2(float16_t const *a) { @@ -13230,7 +13230,7 @@ float16x8x2_t test_vld1q_f16_x2(float16_t const *a) { // CHECK: store { <4 x float>, <4 x float> } [[VLD1XN]], { <4 x float>, <4 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x4x2_t, %struct.float32x4x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.float32x4x2_t [[TMP6]] float32x4x2_t test_vld1q_f32_x2(float32_t const *a) { @@ -13248,7 +13248,7 @@ float32x4x2_t test_vld1q_f32_x2(float32_t const *a) { // CHECK: store { <2 x double>, <2 x double> } [[VLD1XN]], { <2 x double>, <2 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x2x2_t, %struct.float64x2x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.float64x2x2_t [[TMP6]] float64x2x2_t test_vld1q_f64_x2(float64_t const *a) { @@ -13264,7 +13264,7 @@ float64x2x2_t test_vld1q_f64_x2(float64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.poly8x16x2_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x16x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 32, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.poly8x16x2_t, %struct.poly8x16x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly8x16x2_t [[TMP4]] poly8x16x2_t test_vld1q_p8_x2(poly8_t const *a) { @@ -13282,7 +13282,7 @@ poly8x16x2_t test_vld1q_p8_x2(poly8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x8x2_t, %struct.poly16x8x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly16x8x2_t [[TMP6]] poly16x8x2_t test_vld1q_p16_x2(poly16_t const *a) { @@ -13300,7 +13300,7 @@ poly16x8x2_t test_vld1q_p16_x2(poly16_t const *a) { // CHECK: store { <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly64x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly64x2x2_t, %struct.poly64x2x2_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly64x2x2_t [[TMP6]] poly64x2x2_t test_vld1q_p64_x2(poly64_t const *a) { @@ -13316,7 +13316,7 @@ poly64x2x2_t test_vld1q_p64_x2(poly64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.uint8x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 16, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.uint8x8x2_t, %struct.uint8x8x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint8x8x2_t [[TMP4]] uint8x8x2_t test_vld1_u8_x2(uint8_t const *a) { @@ -13334,7 +13334,7 @@ uint8x8x2_t test_vld1_u8_x2(uint8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x4x2_t, %struct.uint16x4x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint16x4x2_t [[TMP6]] uint16x4x2_t test_vld1_u16_x2(uint16_t const *a) { @@ -13352,7 +13352,7 @@ uint16x4x2_t test_vld1_u16_x2(uint16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32> } [[VLD1XN]], { <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x2x2_t, %struct.uint32x2x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint32x2x2_t [[TMP6]] uint32x2x2_t test_vld1_u32_x2(uint32_t const *a) { @@ -13370,7 +13370,7 @@ uint32x2x2_t test_vld1_u32_x2(uint32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x1x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x1x2_t, %struct.uint64x1x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint64x1x2_t [[TMP6]] uint64x1x2_t test_vld1_u64_x2(uint64_t const *a) { @@ -13386,7 +13386,7 @@ uint64x1x2_t test_vld1_u64_x2(uint64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.int8x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 16, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.int8x8x2_t, %struct.int8x8x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.int8x8x2_t [[TMP4]] int8x8x2_t test_vld1_s8_x2(int8_t const *a) { @@ -13404,7 +13404,7 @@ int8x8x2_t test_vld1_s8_x2(int8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x4x2_t, %struct.int16x4x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.int16x4x2_t [[TMP6]] int16x4x2_t test_vld1_s16_x2(int16_t const *a) { @@ -13422,7 +13422,7 @@ int16x4x2_t test_vld1_s16_x2(int16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32> } [[VLD1XN]], { <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x2x2_t, %struct.int32x2x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.int32x2x2_t [[TMP6]] int32x2x2_t test_vld1_s32_x2(int32_t const *a) { @@ -13440,7 +13440,7 @@ int32x2x2_t test_vld1_s32_x2(int32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x1x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x1x2_t, %struct.int64x1x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.int64x1x2_t [[TMP6]] int64x1x2_t test_vld1_s64_x2(int64_t const *a) { @@ -13458,7 +13458,7 @@ int64x1x2_t test_vld1_s64_x2(int64_t const *a) { // CHECK: store { <4 x half>, <4 x half> } [[VLD1XN]], { <4 x half>, <4 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x4x2_t, %struct.float16x4x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.float16x4x2_t [[TMP6]] float16x4x2_t test_vld1_f16_x2(float16_t const *a) { @@ -13476,7 +13476,7 @@ float16x4x2_t test_vld1_f16_x2(float16_t const *a) { // CHECK: store { <2 x float>, <2 x float> } [[VLD1XN]], { <2 x float>, <2 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x2x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x2x2_t, %struct.float32x2x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.float32x2x2_t [[TMP6]] float32x2x2_t test_vld1_f32_x2(float32_t const *a) { @@ -13494,7 +13494,7 @@ float32x2x2_t test_vld1_f32_x2(float32_t const *a) { // CHECK: store { <1 x double>, <1 x double> } [[VLD1XN]], { <1 x double>, <1 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x1x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x1x2_t, %struct.float64x1x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.float64x1x2_t [[TMP6]] float64x1x2_t test_vld1_f64_x2(float64_t const *a) { @@ -13510,7 +13510,7 @@ float64x1x2_t test_vld1_f64_x2(float64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.poly8x8x2_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x8x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 16, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.poly8x8x2_t, %struct.poly8x8x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly8x8x2_t [[TMP4]] poly8x8x2_t test_vld1_p8_x2(poly8_t const *a) { @@ -13528,7 +13528,7 @@ poly8x8x2_t test_vld1_p8_x2(poly8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x4x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x4x2_t, %struct.poly16x4x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly16x4x2_t [[TMP6]] poly16x4x2_t test_vld1_p16_x2(poly16_t const *a) { @@ -13546,7 +13546,7 @@ poly16x4x2_t test_vld1_p16_x2(poly16_t const *a) { // CHECK: store { <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x1x2_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly64x1x2_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 16, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly64x1x2_t, %struct.poly64x1x2_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly64x1x2_t [[TMP6]] poly64x1x2_t test_vld1_p64_x2(poly64_t const *a) { @@ -13562,7 +13562,7 @@ poly64x1x2_t test_vld1_p64_x2(poly64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.uint8x16x3_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x16x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 48, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.uint8x16x3_t, %struct.uint8x16x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint8x16x3_t [[TMP4]] uint8x16x3_t test_vld1q_u8_x3(uint8_t const *a) { @@ -13580,7 +13580,7 @@ uint8x16x3_t test_vld1q_u8_x3(uint8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x8x3_t, %struct.uint16x8x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint16x8x3_t [[TMP6]] uint16x8x3_t test_vld1q_u16_x3(uint16_t const *a) { @@ -13598,7 +13598,7 @@ uint16x8x3_t test_vld1q_u16_x3(uint16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32>, <4 x i32> } [[VLD1XN]], { <4 x i32>, <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x4x3_t, %struct.uint32x4x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint32x4x3_t [[TMP6]] uint32x4x3_t test_vld1q_u32_x3(uint32_t const *a) { @@ -13616,7 +13616,7 @@ uint32x4x3_t test_vld1q_u32_x3(uint32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x2x3_t, %struct.uint64x2x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint64x2x3_t [[TMP6]] uint64x2x3_t test_vld1q_u64_x3(uint64_t const *a) { @@ -13632,7 +13632,7 @@ uint64x2x3_t test_vld1q_u64_x3(uint64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.int8x16x3_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x16x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 48, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.int8x16x3_t, %struct.int8x16x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.int8x16x3_t [[TMP4]] int8x16x3_t test_vld1q_s8_x3(int8_t const *a) { @@ -13650,7 +13650,7 @@ int8x16x3_t test_vld1q_s8_x3(int8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x8x3_t, %struct.int16x8x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.int16x8x3_t [[TMP6]] int16x8x3_t test_vld1q_s16_x3(int16_t const *a) { @@ -13668,7 +13668,7 @@ int16x8x3_t test_vld1q_s16_x3(int16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32>, <4 x i32> } [[VLD1XN]], { <4 x i32>, <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x4x3_t, %struct.int32x4x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.int32x4x3_t [[TMP6]] int32x4x3_t test_vld1q_s32_x3(int32_t const *a) { @@ -13686,7 +13686,7 @@ int32x4x3_t test_vld1q_s32_x3(int32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x2x3_t, %struct.int64x2x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.int64x2x3_t [[TMP6]] int64x2x3_t test_vld1q_s64_x3(int64_t const *a) { @@ -13704,7 +13704,7 @@ int64x2x3_t test_vld1q_s64_x3(int64_t const *a) { // CHECK: store { <8 x half>, <8 x half>, <8 x half> } [[VLD1XN]], { <8 x half>, <8 x half>, <8 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x8x3_t, %struct.float16x8x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.float16x8x3_t [[TMP6]] float16x8x3_t test_vld1q_f16_x3(float16_t const *a) { @@ -13722,7 +13722,7 @@ float16x8x3_t test_vld1q_f16_x3(float16_t const *a) { // CHECK: store { <4 x float>, <4 x float>, <4 x float> } [[VLD1XN]], { <4 x float>, <4 x float>, <4 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x4x3_t, %struct.float32x4x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.float32x4x3_t [[TMP6]] float32x4x3_t test_vld1q_f32_x3(float32_t const *a) { @@ -13740,7 +13740,7 @@ float32x4x3_t test_vld1q_f32_x3(float32_t const *a) { // CHECK: store { <2 x double>, <2 x double>, <2 x double> } [[VLD1XN]], { <2 x double>, <2 x double>, <2 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x2x3_t, %struct.float64x2x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.float64x2x3_t [[TMP6]] float64x2x3_t test_vld1q_f64_x3(float64_t const *a) { @@ -13756,7 +13756,7 @@ float64x2x3_t test_vld1q_f64_x3(float64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.poly8x16x3_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x16x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 48, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.poly8x16x3_t, %struct.poly8x16x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly8x16x3_t [[TMP4]] poly8x16x3_t test_vld1q_p8_x3(poly8_t const *a) { @@ -13774,7 +13774,7 @@ poly8x16x3_t test_vld1q_p8_x3(poly8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x8x3_t, %struct.poly16x8x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly16x8x3_t [[TMP6]] poly16x8x3_t test_vld1q_p16_x3(poly16_t const *a) { @@ -13792,7 +13792,7 @@ poly16x8x3_t test_vld1q_p16_x3(poly16_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly64x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 48, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly64x2x3_t, %struct.poly64x2x3_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly64x2x3_t [[TMP6]] poly64x2x3_t test_vld1q_p64_x3(poly64_t const *a) { @@ -13808,7 +13808,7 @@ poly64x2x3_t test_vld1q_p64_x3(poly64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.uint8x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 24, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.uint8x8x3_t, %struct.uint8x8x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint8x8x3_t [[TMP4]] uint8x8x3_t test_vld1_u8_x3(uint8_t const *a) { @@ -13826,7 +13826,7 @@ uint8x8x3_t test_vld1_u8_x3(uint8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x4x3_t, %struct.uint16x4x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint16x4x3_t [[TMP6]] uint16x4x3_t test_vld1_u16_x3(uint16_t const *a) { @@ -13844,7 +13844,7 @@ uint16x4x3_t test_vld1_u16_x3(uint16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32>, <2 x i32> } [[VLD1XN]], { <2 x i32>, <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x2x3_t, %struct.uint32x2x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint32x2x3_t [[TMP6]] uint32x2x3_t test_vld1_u32_x3(uint32_t const *a) { @@ -13862,7 +13862,7 @@ uint32x2x3_t test_vld1_u32_x3(uint32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x1x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x1x3_t, %struct.uint64x1x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint64x1x3_t [[TMP6]] uint64x1x3_t test_vld1_u64_x3(uint64_t const *a) { @@ -13878,7 +13878,7 @@ uint64x1x3_t test_vld1_u64_x3(uint64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.int8x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 24, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.int8x8x3_t, %struct.int8x8x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.int8x8x3_t [[TMP4]] int8x8x3_t test_vld1_s8_x3(int8_t const *a) { @@ -13896,7 +13896,7 @@ int8x8x3_t test_vld1_s8_x3(int8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x4x3_t, %struct.int16x4x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.int16x4x3_t [[TMP6]] int16x4x3_t test_vld1_s16_x3(int16_t const *a) { @@ -13914,7 +13914,7 @@ int16x4x3_t test_vld1_s16_x3(int16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32>, <2 x i32> } [[VLD1XN]], { <2 x i32>, <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x2x3_t, %struct.int32x2x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.int32x2x3_t [[TMP6]] int32x2x3_t test_vld1_s32_x3(int32_t const *a) { @@ -13932,7 +13932,7 @@ int32x2x3_t test_vld1_s32_x3(int32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x1x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x1x3_t, %struct.int64x1x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.int64x1x3_t [[TMP6]] int64x1x3_t test_vld1_s64_x3(int64_t const *a) { @@ -13950,7 +13950,7 @@ int64x1x3_t test_vld1_s64_x3(int64_t const *a) { // CHECK: store { <4 x half>, <4 x half>, <4 x half> } [[VLD1XN]], { <4 x half>, <4 x half>, <4 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x4x3_t, %struct.float16x4x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.float16x4x3_t [[TMP6]] float16x4x3_t test_vld1_f16_x3(float16_t const *a) { @@ -13968,7 +13968,7 @@ float16x4x3_t test_vld1_f16_x3(float16_t const *a) { // CHECK: store { <2 x float>, <2 x float>, <2 x float> } [[VLD1XN]], { <2 x float>, <2 x float>, <2 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x2x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x2x3_t, %struct.float32x2x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.float32x2x3_t [[TMP6]] float32x2x3_t test_vld1_f32_x3(float32_t const *a) { @@ -13986,7 +13986,7 @@ float32x2x3_t test_vld1_f32_x3(float32_t const *a) { // CHECK: store { <1 x double>, <1 x double>, <1 x double> } [[VLD1XN]], { <1 x double>, <1 x double>, <1 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x1x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x1x3_t, %struct.float64x1x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.float64x1x3_t [[TMP6]] float64x1x3_t test_vld1_f64_x3(float64_t const *a) { @@ -14002,7 +14002,7 @@ float64x1x3_t test_vld1_f64_x3(float64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.poly8x8x3_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x8x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 24, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.poly8x8x3_t, %struct.poly8x8x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly8x8x3_t [[TMP4]] poly8x8x3_t test_vld1_p8_x3(poly8_t const *a) { @@ -14020,7 +14020,7 @@ poly8x8x3_t test_vld1_p8_x3(poly8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x4x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x4x3_t, %struct.poly16x4x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly16x4x3_t [[TMP6]] poly16x4x3_t test_vld1_p16_x3(poly16_t const *a) { @@ -14038,7 +14038,7 @@ poly16x4x3_t test_vld1_p16_x3(poly16_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x1x3_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly64x1x3_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 24, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly64x1x3_t, %struct.poly64x1x3_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly64x1x3_t [[TMP6]] poly64x1x3_t test_vld1_p64_x3(poly64_t const *a) { @@ -14054,7 +14054,7 @@ poly64x1x3_t test_vld1_p64_x3(poly64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.uint8x16x4_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x16x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 64, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.uint8x16x4_t, %struct.uint8x16x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint8x16x4_t [[TMP4]] uint8x16x4_t test_vld1q_u8_x4(uint8_t const *a) { @@ -14072,7 +14072,7 @@ uint8x16x4_t test_vld1q_u8_x4(uint8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x8x4_t, %struct.uint16x8x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint16x8x4_t [[TMP6]] uint16x8x4_t test_vld1q_u16_x4(uint16_t const *a) { @@ -14090,7 +14090,7 @@ uint16x8x4_t test_vld1q_u16_x4(uint16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> } [[VLD1XN]], { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x4x4_t, %struct.uint32x4x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint32x4x4_t [[TMP6]] uint32x4x4_t test_vld1q_u32_x4(uint32_t const *a) { @@ -14108,7 +14108,7 @@ uint32x4x4_t test_vld1q_u32_x4(uint32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x2x4_t, %struct.uint64x2x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.uint64x2x4_t [[TMP6]] uint64x2x4_t test_vld1q_u64_x4(uint64_t const *a) { @@ -14124,7 +14124,7 @@ uint64x2x4_t test_vld1q_u64_x4(uint64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.int8x16x4_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x16x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 64, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.int8x16x4_t, %struct.int8x16x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.int8x16x4_t [[TMP4]] int8x16x4_t test_vld1q_s8_x4(int8_t const *a) { @@ -14142,7 +14142,7 @@ int8x16x4_t test_vld1q_s8_x4(int8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x8x4_t, %struct.int16x8x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.int16x8x4_t [[TMP6]] int16x8x4_t test_vld1q_s16_x4(int16_t const *a) { @@ -14160,7 +14160,7 @@ int16x8x4_t test_vld1q_s16_x4(int16_t const *a) { // CHECK: store { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> } [[VLD1XN]], { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x4x4_t, %struct.int32x4x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.int32x4x4_t [[TMP6]] int32x4x4_t test_vld1q_s32_x4(int32_t const *a) { @@ -14178,7 +14178,7 @@ int32x4x4_t test_vld1q_s32_x4(int32_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x2x4_t, %struct.int64x2x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.int64x2x4_t [[TMP6]] int64x2x4_t test_vld1q_s64_x4(int64_t const *a) { @@ -14196,7 +14196,7 @@ int64x2x4_t test_vld1q_s64_x4(int64_t const *a) { // CHECK: store { <8 x half>, <8 x half>, <8 x half>, <8 x half> } [[VLD1XN]], { <8 x half>, <8 x half>, <8 x half>, <8 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x8x4_t, %struct.float16x8x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.float16x8x4_t [[TMP6]] float16x8x4_t test_vld1q_f16_x4(float16_t const *a) { @@ -14214,7 +14214,7 @@ float16x8x4_t test_vld1q_f16_x4(float16_t const *a) { // CHECK: store { <4 x float>, <4 x float>, <4 x float>, <4 x float> } [[VLD1XN]], { <4 x float>, <4 x float>, <4 x float>, <4 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x4x4_t, %struct.float32x4x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.float32x4x4_t [[TMP6]] float32x4x4_t test_vld1q_f32_x4(float32_t const *a) { @@ -14232,7 +14232,7 @@ float32x4x4_t test_vld1q_f32_x4(float32_t const *a) { // CHECK: store { <2 x double>, <2 x double>, <2 x double>, <2 x double> } [[VLD1XN]], { <2 x double>, <2 x double>, <2 x double>, <2 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x2x4_t, %struct.float64x2x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.float64x2x4_t [[TMP6]] float64x2x4_t test_vld1q_f64_x4(float64_t const *a) { @@ -14248,7 +14248,7 @@ float64x2x4_t test_vld1q_f64_x4(float64_t const *a) { // CHECK: store { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } [[VLD1XN]], { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.poly8x16x4_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x16x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP2]], i8* align 16 [[TMP3]], i64 64, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.poly8x16x4_t, %struct.poly8x16x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly8x16x4_t [[TMP4]] poly8x16x4_t test_vld1q_p8_x4(poly8_t const *a) { @@ -14266,7 +14266,7 @@ poly8x16x4_t test_vld1q_p8_x4(poly8_t const *a) { // CHECK: store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x8x4_t, %struct.poly16x8x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly16x8x4_t [[TMP6]] poly16x8x4_t test_vld1q_p16_x4(poly16_t const *a) { @@ -14284,7 +14284,7 @@ poly16x8x4_t test_vld1q_p16_x4(poly16_t const *a) { // CHECK: store { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> } [[VLD1XN]], { <2 x i64>, <2 x i64>, <2 x i64>, <2 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly64x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP4]], i8* align 16 [[TMP5]], i64 64, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly64x2x4_t, %struct.poly64x2x4_t* [[RETVAL]], align 16 // CHECK: ret %struct.poly64x2x4_t [[TMP6]] poly64x2x4_t test_vld1q_p64_x4(poly64_t const *a) { @@ -14300,7 +14300,7 @@ poly64x2x4_t test_vld1q_p64_x4(poly64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.uint8x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.uint8x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 32, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.uint8x8x4_t, %struct.uint8x8x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint8x8x4_t [[TMP4]] uint8x8x4_t test_vld1_u8_x4(uint8_t const *a) { @@ -14318,7 +14318,7 @@ uint8x8x4_t test_vld1_u8_x4(uint8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint16x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint16x4x4_t, %struct.uint16x4x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint16x4x4_t [[TMP6]] uint16x4x4_t test_vld1_u16_x4(uint16_t const *a) { @@ -14336,7 +14336,7 @@ uint16x4x4_t test_vld1_u16_x4(uint16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } [[VLD1XN]], { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint32x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint32x2x4_t, %struct.uint32x2x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint32x2x4_t [[TMP6]] uint32x2x4_t test_vld1_u32_x4(uint32_t const *a) { @@ -14354,7 +14354,7 @@ uint32x2x4_t test_vld1_u32_x4(uint32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.uint64x1x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.uint64x1x4_t, %struct.uint64x1x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.uint64x1x4_t [[TMP6]] uint64x1x4_t test_vld1_u64_x4(uint64_t const *a) { @@ -14370,7 +14370,7 @@ uint64x1x4_t test_vld1_u64_x4(uint64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.int8x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.int8x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 32, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.int8x8x4_t, %struct.int8x8x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.int8x8x4_t [[TMP4]] int8x8x4_t test_vld1_s8_x4(int8_t const *a) { @@ -14388,7 +14388,7 @@ int8x8x4_t test_vld1_s8_x4(int8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int16x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int16x4x4_t, %struct.int16x4x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.int16x4x4_t [[TMP6]] int16x4x4_t test_vld1_s16_x4(int16_t const *a) { @@ -14406,7 +14406,7 @@ int16x4x4_t test_vld1_s16_x4(int16_t const *a) { // CHECK: store { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } [[VLD1XN]], { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int32x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int32x2x4_t, %struct.int32x2x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.int32x2x4_t [[TMP6]] int32x2x4_t test_vld1_s32_x4(int32_t const *a) { @@ -14424,7 +14424,7 @@ int32x2x4_t test_vld1_s32_x4(int32_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.int64x1x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.int64x1x4_t, %struct.int64x1x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.int64x1x4_t [[TMP6]] int64x1x4_t test_vld1_s64_x4(int64_t const *a) { @@ -14442,7 +14442,7 @@ int64x1x4_t test_vld1_s64_x4(int64_t const *a) { // CHECK: store { <4 x half>, <4 x half>, <4 x half>, <4 x half> } [[VLD1XN]], { <4 x half>, <4 x half>, <4 x half>, <4 x half> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float16x4x4_t, %struct.float16x4x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.float16x4x4_t [[TMP6]] float16x4x4_t test_vld1_f16_x4(float16_t const *a) { @@ -14460,7 +14460,7 @@ float16x4x4_t test_vld1_f16_x4(float16_t const *a) { // CHECK: store { <2 x float>, <2 x float>, <2 x float>, <2 x float> } [[VLD1XN]], { <2 x float>, <2 x float>, <2 x float>, <2 x float> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float32x2x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float32x2x4_t, %struct.float32x2x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.float32x2x4_t [[TMP6]] float32x2x4_t test_vld1_f32_x4(float32_t const *a) { @@ -14478,7 +14478,7 @@ float32x2x4_t test_vld1_f32_x4(float32_t const *a) { // CHECK: store { <1 x double>, <1 x double>, <1 x double>, <1 x double> } [[VLD1XN]], { <1 x double>, <1 x double>, <1 x double>, <1 x double> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.float64x1x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.float64x1x4_t, %struct.float64x1x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.float64x1x4_t [[TMP6]] float64x1x4_t test_vld1_f64_x4(float64_t const *a) { @@ -14494,7 +14494,7 @@ float64x1x4_t test_vld1_f64_x4(float64_t const *a) { // CHECK: store { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } [[VLD1XN]], { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> }* [[TMP1]] // CHECK: [[TMP2:%.*]] = bitcast %struct.poly8x8x4_t* [[RETVAL]] to i8* // CHECK: [[TMP3:%.*]] = bitcast %struct.poly8x8x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP2]], i8* [[TMP3]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP2]], i8* align 8 [[TMP3]], i64 32, i1 false) // CHECK: [[TMP4:%.*]] = load %struct.poly8x8x4_t, %struct.poly8x8x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly8x8x4_t [[TMP4]] poly8x8x4_t test_vld1_p8_x4(poly8_t const *a) { @@ -14512,7 +14512,7 @@ poly8x8x4_t test_vld1_p8_x4(poly8_t const *a) { // CHECK: store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly16x4x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly16x4x4_t, %struct.poly16x4x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly16x4x4_t [[TMP6]] poly16x4x4_t test_vld1_p16_x4(poly16_t const *a) { @@ -14530,7 +14530,7 @@ poly16x4x4_t test_vld1_p16_x4(poly16_t const *a) { // CHECK: store { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> } [[VLD1XN]], { <1 x i64>, <1 x i64>, <1 x i64>, <1 x i64> }* [[TMP3]] // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x1x4_t* [[RETVAL]] to i8* // CHECK: [[TMP5:%.*]] = bitcast %struct.poly64x1x4_t* [[__RET]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP4]], i8* align 8 [[TMP5]], i64 32, i1 false) // CHECK: [[TMP6:%.*]] = load %struct.poly64x1x4_t, %struct.poly64x1x4_t* [[RETVAL]], align 8 // CHECK: ret %struct.poly64x1x4_t [[TMP6]] poly64x1x4_t test_vld1_p64_x4(poly64_t const *a) { @@ -14544,7 +14544,7 @@ poly64x1x4_t test_vld1_p64_x4(poly64_t const *a) { // CHECK: store [2 x <16 x i8>] [[B]].coerce, [2 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x16x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x16x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x16x2_t, %struct.uint8x16x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <16 x i8>], [2 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -14564,7 +14564,7 @@ void test_vst1q_u8_x2(uint8_t *a, uint8x16x2_t b) { // CHECK: store [2 x <8 x i16>] [[B]].coerce, [2 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x8x2_t, %struct.uint16x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i16>], [2 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -14590,7 +14590,7 @@ void test_vst1q_u16_x2(uint16_t *a, uint16x8x2_t b) { // CHECK: store [2 x <4 x i32>] [[B]].coerce, [2 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x4x2_t, %struct.uint32x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i32>], [2 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -14616,7 +14616,7 @@ void test_vst1q_u32_x2(uint32_t *a, uint32x4x2_t b) { // CHECK: store [2 x <2 x i64>] [[B]].coerce, [2 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x2x2_t, %struct.uint64x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i64>], [2 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -14642,7 +14642,7 @@ void test_vst1q_u64_x2(uint64_t *a, uint64x2x2_t b) { // CHECK: store [2 x <16 x i8>] [[B]].coerce, [2 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x16x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x16x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x16x2_t, %struct.int8x16x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <16 x i8>], [2 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -14662,7 +14662,7 @@ void test_vst1q_s8_x2(int8_t *a, int8x16x2_t b) { // CHECK: store [2 x <8 x i16>] [[B]].coerce, [2 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x8x2_t, %struct.int16x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i16>], [2 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -14688,7 +14688,7 @@ void test_vst1q_s16_x2(int16_t *a, int16x8x2_t b) { // CHECK: store [2 x <4 x i32>] [[B]].coerce, [2 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x4x2_t, %struct.int32x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i32>], [2 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -14714,7 +14714,7 @@ void test_vst1q_s32_x2(int32_t *a, int32x4x2_t b) { // CHECK: store [2 x <2 x i64>] [[B]].coerce, [2 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x2x2_t, %struct.int64x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i64>], [2 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -14740,7 +14740,7 @@ void test_vst1q_s64_x2(int64_t *a, int64x2x2_t b) { // CHECK: store [2 x <8 x half>] [[B]].coerce, [2 x <8 x half>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x8x2_t, %struct.float16x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL]], i64 0, i64 0 @@ -14766,7 +14766,7 @@ void test_vst1q_f16_x2(float16_t *a, float16x8x2_t b) { // CHECK: store [2 x <4 x float>] [[B]].coerce, [2 x <4 x float>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x4x2_t, %struct.float32x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x float>], [2 x <4 x float>]* [[VAL]], i64 0, i64 0 @@ -14792,7 +14792,7 @@ void test_vst1q_f32_x2(float32_t *a, float32x4x2_t b) { // CHECK: store [2 x <2 x double>] [[B]].coerce, [2 x <2 x double>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x2x2_t, %struct.float64x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x double>], [2 x <2 x double>]* [[VAL]], i64 0, i64 0 @@ -14818,7 +14818,7 @@ void test_vst1q_f64_x2(float64_t *a, float64x2x2_t b) { // CHECK: store [2 x <16 x i8>] [[B]].coerce, [2 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x16x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x16x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x16x2_t, %struct.poly8x16x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <16 x i8>], [2 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -14838,7 +14838,7 @@ void test_vst1q_p8_x2(poly8_t *a, poly8x16x2_t b) { // CHECK: store [2 x <8 x i16>] [[B]].coerce, [2 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x8x2_t, %struct.poly16x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i16>], [2 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -14864,7 +14864,7 @@ void test_vst1q_p16_x2(poly16_t *a, poly16x8x2_t b) { // CHECK: store [2 x <2 x i64>] [[B]].coerce, [2 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly64x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly64x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly64x2x2_t, %struct.poly64x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i64>], [2 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -14890,7 +14890,7 @@ void test_vst1q_p64_x2(poly64_t *a, poly64x2x2_t b) { // CHECK: store [2 x <8 x i8>] [[B]].coerce, [2 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x8x2_t, %struct.uint8x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i8>], [2 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -14910,7 +14910,7 @@ void test_vst1_u8_x2(uint8_t *a, uint8x8x2_t b) { // CHECK: store [2 x <4 x i16>] [[B]].coerce, [2 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x4x2_t, %struct.uint16x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i16>], [2 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -14936,7 +14936,7 @@ void test_vst1_u16_x2(uint16_t *a, uint16x4x2_t b) { // CHECK: store [2 x <2 x i32>] [[B]].coerce, [2 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x2x2_t, %struct.uint32x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i32>], [2 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -14962,7 +14962,7 @@ void test_vst1_u32_x2(uint32_t *a, uint32x2x2_t b) { // CHECK: store [2 x <1 x i64>] [[B]].coerce, [2 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x1x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x1x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x1x2_t, %struct.uint64x1x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <1 x i64>], [2 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -14988,7 +14988,7 @@ void test_vst1_u64_x2(uint64_t *a, uint64x1x2_t b) { // CHECK: store [2 x <8 x i8>] [[B]].coerce, [2 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x8x2_t, %struct.int8x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i8>], [2 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -15008,7 +15008,7 @@ void test_vst1_s8_x2(int8_t *a, int8x8x2_t b) { // CHECK: store [2 x <4 x i16>] [[B]].coerce, [2 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x4x2_t, %struct.int16x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i16>], [2 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -15034,7 +15034,7 @@ void test_vst1_s16_x2(int16_t *a, int16x4x2_t b) { // CHECK: store [2 x <2 x i32>] [[B]].coerce, [2 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x2x2_t, %struct.int32x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x i32>], [2 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -15060,7 +15060,7 @@ void test_vst1_s32_x2(int32_t *a, int32x2x2_t b) { // CHECK: store [2 x <1 x i64>] [[B]].coerce, [2 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x1x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x1x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x1x2_t, %struct.int64x1x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <1 x i64>], [2 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -15086,7 +15086,7 @@ void test_vst1_s64_x2(int64_t *a, int64x1x2_t b) { // CHECK: store [2 x <4 x half>] [[B]].coerce, [2 x <4 x half>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x4x2_t, %struct.float16x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL]], i64 0, i64 0 @@ -15112,7 +15112,7 @@ void test_vst1_f16_x2(float16_t *a, float16x4x2_t b) { // CHECK: store [2 x <2 x float>] [[B]].coerce, [2 x <2 x float>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x2x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x2x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x2x2_t, %struct.float32x2x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <2 x float>], [2 x <2 x float>]* [[VAL]], i64 0, i64 0 @@ -15138,7 +15138,7 @@ void test_vst1_f32_x2(float32_t *a, float32x2x2_t b) { // CHECK: store [2 x <1 x double>] [[B]].coerce, [2 x <1 x double>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x1x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x1x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x1x2_t, %struct.float64x1x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <1 x double>], [2 x <1 x double>]* [[VAL]], i64 0, i64 0 @@ -15164,7 +15164,7 @@ void test_vst1_f64_x2(float64_t *a, float64x1x2_t b) { // CHECK: store [2 x <8 x i8>] [[B]].coerce, [2 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x8x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x8x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x8x2_t, %struct.poly8x8x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <8 x i8>], [2 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -15184,7 +15184,7 @@ void test_vst1_p8_x2(poly8_t *a, poly8x8x2_t b) { // CHECK: store [2 x <4 x i16>] [[B]].coerce, [2 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x4x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x4x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x4x2_t, %struct.poly16x4x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <4 x i16>], [2 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -15210,7 +15210,7 @@ void test_vst1_p16_x2(poly16_t *a, poly16x4x2_t b) { // CHECK: store [2 x <1 x i64>] [[B]].coerce, [2 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly64x1x2_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly64x1x2_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 16, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 16, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly64x1x2_t, %struct.poly64x1x2_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x <1 x i64>], [2 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -15236,7 +15236,7 @@ void test_vst1_p64_x2(poly64_t *a, poly64x1x2_t b) { // CHECK: store [3 x <16 x i8>] [[B]].coerce, [3 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x16x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x16x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x16x3_t, %struct.uint8x16x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <16 x i8>], [3 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -15259,7 +15259,7 @@ void test_vst1q_u8_x3(uint8_t *a, uint8x16x3_t b) { // CHECK: store [3 x <8 x i16>] [[B]].coerce, [3 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x8x3_t, %struct.uint16x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i16>], [3 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -15290,7 +15290,7 @@ void test_vst1q_u16_x3(uint16_t *a, uint16x8x3_t b) { // CHECK: store [3 x <4 x i32>] [[B]].coerce, [3 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x4x3_t, %struct.uint32x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i32>], [3 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -15321,7 +15321,7 @@ void test_vst1q_u32_x3(uint32_t *a, uint32x4x3_t b) { // CHECK: store [3 x <2 x i64>] [[B]].coerce, [3 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x2x3_t, %struct.uint64x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i64>], [3 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -15352,7 +15352,7 @@ void test_vst1q_u64_x3(uint64_t *a, uint64x2x3_t b) { // CHECK: store [3 x <16 x i8>] [[B]].coerce, [3 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x16x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x16x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x16x3_t, %struct.int8x16x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <16 x i8>], [3 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -15375,7 +15375,7 @@ void test_vst1q_s8_x3(int8_t *a, int8x16x3_t b) { // CHECK: store [3 x <8 x i16>] [[B]].coerce, [3 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x8x3_t, %struct.int16x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i16>], [3 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -15406,7 +15406,7 @@ void test_vst1q_s16_x3(int16_t *a, int16x8x3_t b) { // CHECK: store [3 x <4 x i32>] [[B]].coerce, [3 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x4x3_t, %struct.int32x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i32>], [3 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -15437,7 +15437,7 @@ void test_vst1q_s32_x3(int32_t *a, int32x4x3_t b) { // CHECK: store [3 x <2 x i64>] [[B]].coerce, [3 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x2x3_t, %struct.int64x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i64>], [3 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -15468,7 +15468,7 @@ void test_vst1q_s64_x3(int64_t *a, int64x2x3_t b) { // CHECK: store [3 x <8 x half>] [[B]].coerce, [3 x <8 x half>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x8x3_t, %struct.float16x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL]], i64 0, i64 0 @@ -15499,7 +15499,7 @@ void test_vst1q_f16_x3(float16_t *a, float16x8x3_t b) { // CHECK: store [3 x <4 x float>] [[B]].coerce, [3 x <4 x float>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x4x3_t, %struct.float32x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x float>], [3 x <4 x float>]* [[VAL]], i64 0, i64 0 @@ -15530,7 +15530,7 @@ void test_vst1q_f32_x3(float32_t *a, float32x4x3_t b) { // CHECK: store [3 x <2 x double>] [[B]].coerce, [3 x <2 x double>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x2x3_t, %struct.float64x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x double>], [3 x <2 x double>]* [[VAL]], i64 0, i64 0 @@ -15561,7 +15561,7 @@ void test_vst1q_f64_x3(float64_t *a, float64x2x3_t b) { // CHECK: store [3 x <16 x i8>] [[B]].coerce, [3 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x16x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x16x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x16x3_t, %struct.poly8x16x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <16 x i8>], [3 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -15584,7 +15584,7 @@ void test_vst1q_p8_x3(poly8_t *a, poly8x16x3_t b) { // CHECK: store [3 x <8 x i16>] [[B]].coerce, [3 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x8x3_t, %struct.poly16x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i16>], [3 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -15615,7 +15615,7 @@ void test_vst1q_p16_x3(poly16_t *a, poly16x8x3_t b) { // CHECK: store [3 x <2 x i64>] [[B]].coerce, [3 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly64x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly64x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 48, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 48, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly64x2x3_t, %struct.poly64x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i64>], [3 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -15646,7 +15646,7 @@ void test_vst1q_p64_x3(poly64_t *a, poly64x2x3_t b) { // CHECK: store [3 x <8 x i8>] [[B]].coerce, [3 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x8x3_t, %struct.uint8x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i8>], [3 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -15669,7 +15669,7 @@ void test_vst1_u8_x3(uint8_t *a, uint8x8x3_t b) { // CHECK: store [3 x <4 x i16>] [[B]].coerce, [3 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x4x3_t, %struct.uint16x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i16>], [3 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -15700,7 +15700,7 @@ void test_vst1_u16_x3(uint16_t *a, uint16x4x3_t b) { // CHECK: store [3 x <2 x i32>] [[B]].coerce, [3 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x2x3_t, %struct.uint32x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i32>], [3 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -15731,7 +15731,7 @@ void test_vst1_u32_x3(uint32_t *a, uint32x2x3_t b) { // CHECK: store [3 x <1 x i64>] [[B]].coerce, [3 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x1x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x1x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x1x3_t, %struct.uint64x1x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <1 x i64>], [3 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -15762,7 +15762,7 @@ void test_vst1_u64_x3(uint64_t *a, uint64x1x3_t b) { // CHECK: store [3 x <8 x i8>] [[B]].coerce, [3 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x8x3_t, %struct.int8x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i8>], [3 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -15785,7 +15785,7 @@ void test_vst1_s8_x3(int8_t *a, int8x8x3_t b) { // CHECK: store [3 x <4 x i16>] [[B]].coerce, [3 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x4x3_t, %struct.int16x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i16>], [3 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -15816,7 +15816,7 @@ void test_vst1_s16_x3(int16_t *a, int16x4x3_t b) { // CHECK: store [3 x <2 x i32>] [[B]].coerce, [3 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x2x3_t, %struct.int32x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x i32>], [3 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -15847,7 +15847,7 @@ void test_vst1_s32_x3(int32_t *a, int32x2x3_t b) { // CHECK: store [3 x <1 x i64>] [[B]].coerce, [3 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x1x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x1x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x1x3_t, %struct.int64x1x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <1 x i64>], [3 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -15878,7 +15878,7 @@ void test_vst1_s64_x3(int64_t *a, int64x1x3_t b) { // CHECK: store [3 x <4 x half>] [[B]].coerce, [3 x <4 x half>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x4x3_t, %struct.float16x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL]], i64 0, i64 0 @@ -15909,7 +15909,7 @@ void test_vst1_f16_x3(float16_t *a, float16x4x3_t b) { // CHECK: store [3 x <2 x float>] [[B]].coerce, [3 x <2 x float>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x2x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x2x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x2x3_t, %struct.float32x2x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <2 x float>], [3 x <2 x float>]* [[VAL]], i64 0, i64 0 @@ -15940,7 +15940,7 @@ void test_vst1_f32_x3(float32_t *a, float32x2x3_t b) { // CHECK: store [3 x <1 x double>] [[B]].coerce, [3 x <1 x double>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x1x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x1x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x1x3_t, %struct.float64x1x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <1 x double>], [3 x <1 x double>]* [[VAL]], i64 0, i64 0 @@ -15971,7 +15971,7 @@ void test_vst1_f64_x3(float64_t *a, float64x1x3_t b) { // CHECK: store [3 x <8 x i8>] [[B]].coerce, [3 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x8x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x8x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x8x3_t, %struct.poly8x8x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <8 x i8>], [3 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -15994,7 +15994,7 @@ void test_vst1_p8_x3(poly8_t *a, poly8x8x3_t b) { // CHECK: store [3 x <4 x i16>] [[B]].coerce, [3 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x4x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x4x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x4x3_t, %struct.poly16x4x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <4 x i16>], [3 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -16025,7 +16025,7 @@ void test_vst1_p16_x3(poly16_t *a, poly16x4x3_t b) { // CHECK: store [3 x <1 x i64>] [[B]].coerce, [3 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly64x1x3_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly64x1x3_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 24, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 24, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly64x1x3_t, %struct.poly64x1x3_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [3 x <1 x i64>], [3 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -16056,7 +16056,7 @@ void test_vst1_p64_x3(poly64_t *a, poly64x1x3_t b) { // CHECK: store [4 x <16 x i8>] [[B]].coerce, [4 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x16x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x16x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x16x4_t, %struct.uint8x16x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <16 x i8>], [4 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -16082,7 +16082,7 @@ void test_vst1q_u8_x4(uint8_t *a, uint8x16x4_t b) { // CHECK: store [4 x <8 x i16>] [[B]].coerce, [4 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x8x4_t, %struct.uint16x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i16>], [4 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -16118,7 +16118,7 @@ void test_vst1q_u16_x4(uint16_t *a, uint16x8x4_t b) { // CHECK: store [4 x <4 x i32>] [[B]].coerce, [4 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x4x4_t, %struct.uint32x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i32>], [4 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -16154,7 +16154,7 @@ void test_vst1q_u32_x4(uint32_t *a, uint32x4x4_t b) { // CHECK: store [4 x <2 x i64>] [[B]].coerce, [4 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x2x4_t, %struct.uint64x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i64>], [4 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -16190,7 +16190,7 @@ void test_vst1q_u64_x4(uint64_t *a, uint64x2x4_t b) { // CHECK: store [4 x <16 x i8>] [[B]].coerce, [4 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x16x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x16x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x16x4_t, %struct.int8x16x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <16 x i8>], [4 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -16216,7 +16216,7 @@ void test_vst1q_s8_x4(int8_t *a, int8x16x4_t b) { // CHECK: store [4 x <8 x i16>] [[B]].coerce, [4 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x8x4_t, %struct.int16x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i16>], [4 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -16252,7 +16252,7 @@ void test_vst1q_s16_x4(int16_t *a, int16x8x4_t b) { // CHECK: store [4 x <4 x i32>] [[B]].coerce, [4 x <4 x i32>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x4x4_t, %struct.int32x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i32>], [4 x <4 x i32>]* [[VAL]], i64 0, i64 0 @@ -16288,7 +16288,7 @@ void test_vst1q_s32_x4(int32_t *a, int32x4x4_t b) { // CHECK: store [4 x <2 x i64>] [[B]].coerce, [4 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x2x4_t, %struct.int64x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i64>], [4 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -16324,7 +16324,7 @@ void test_vst1q_s64_x4(int64_t *a, int64x2x4_t b) { // CHECK: store [4 x <8 x half>] [[B]].coerce, [4 x <8 x half>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x8x4_t, %struct.float16x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL]], i64 0, i64 0 @@ -16360,7 +16360,7 @@ void test_vst1q_f16_x4(float16_t *a, float16x8x4_t b) { // CHECK: store [4 x <4 x float>] [[B]].coerce, [4 x <4 x float>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x4x4_t, %struct.float32x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x float>], [4 x <4 x float>]* [[VAL]], i64 0, i64 0 @@ -16396,7 +16396,7 @@ void test_vst1q_f32_x4(float32_t *a, float32x4x4_t b) { // CHECK: store [4 x <2 x double>] [[B]].coerce, [4 x <2 x double>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x2x4_t, %struct.float64x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x double>], [4 x <2 x double>]* [[VAL]], i64 0, i64 0 @@ -16432,7 +16432,7 @@ void test_vst1q_f64_x4(float64_t *a, float64x2x4_t b) { // CHECK: store [4 x <16 x i8>] [[B]].coerce, [4 x <16 x i8>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x16x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x16x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x16x4_t, %struct.poly8x16x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <16 x i8>], [4 x <16 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX]], align 16 @@ -16458,7 +16458,7 @@ void test_vst1q_p8_x4(poly8_t *a, poly8x16x4_t b) { // CHECK: store [4 x <8 x i16>] [[B]].coerce, [4 x <8 x i16>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x8x4_t, %struct.poly16x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i16>], [4 x <8 x i16>]* [[VAL]], i64 0, i64 0 @@ -16494,7 +16494,7 @@ void test_vst1q_p16_x4(poly16_t *a, poly16x8x4_t b) { // CHECK: store [4 x <2 x i64>] [[B]].coerce, [4 x <2 x i64>]* [[COERCE_DIVE]], align 16 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly64x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly64x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 64, i32 16, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[TMP0]], i8* align 16 [[TMP1]], i64 64, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly64x2x4_t, %struct.poly64x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i64>], [4 x <2 x i64>]* [[VAL]], i64 0, i64 0 @@ -16530,7 +16530,7 @@ void test_vst1q_p64_x4(poly64_t *a, poly64x2x4_t b) { // CHECK: store [4 x <8 x i8>] [[B]].coerce, [4 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint8x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint8x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint8x8x4_t, %struct.uint8x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i8>], [4 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -16556,7 +16556,7 @@ void test_vst1_u8_x4(uint8_t *a, uint8x8x4_t b) { // CHECK: store [4 x <4 x i16>] [[B]].coerce, [4 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint16x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint16x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint16x4x4_t, %struct.uint16x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i16>], [4 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -16592,7 +16592,7 @@ void test_vst1_u16_x4(uint16_t *a, uint16x4x4_t b) { // CHECK: store [4 x <2 x i32>] [[B]].coerce, [4 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint32x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint32x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint32x2x4_t, %struct.uint32x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i32>], [4 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -16628,7 +16628,7 @@ void test_vst1_u32_x4(uint32_t *a, uint32x2x4_t b) { // CHECK: store [4 x <1 x i64>] [[B]].coerce, [4 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.uint64x1x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.uint64x1x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.uint64x1x4_t, %struct.uint64x1x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <1 x i64>], [4 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -16664,7 +16664,7 @@ void test_vst1_u64_x4(uint64_t *a, uint64x1x4_t b) { // CHECK: store [4 x <8 x i8>] [[B]].coerce, [4 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int8x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int8x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int8x8x4_t, %struct.int8x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i8>], [4 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -16690,7 +16690,7 @@ void test_vst1_s8_x4(int8_t *a, int8x8x4_t b) { // CHECK: store [4 x <4 x i16>] [[B]].coerce, [4 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int16x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int16x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int16x4x4_t, %struct.int16x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i16>], [4 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -16726,7 +16726,7 @@ void test_vst1_s16_x4(int16_t *a, int16x4x4_t b) { // CHECK: store [4 x <2 x i32>] [[B]].coerce, [4 x <2 x i32>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int32x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int32x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i32* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int32x2x4_t, %struct.int32x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x i32>], [4 x <2 x i32>]* [[VAL]], i64 0, i64 0 @@ -16762,7 +16762,7 @@ void test_vst1_s32_x4(int32_t *a, int32x2x4_t b) { // CHECK: store [4 x <1 x i64>] [[B]].coerce, [4 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.int64x1x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.int64x1x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.int64x1x4_t, %struct.int64x1x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <1 x i64>], [4 x <1 x i64>]* [[VAL]], i64 0, i64 0 @@ -16798,7 +16798,7 @@ void test_vst1_s64_x4(int64_t *a, int64x1x4_t b) { // CHECK: store [4 x <4 x half>] [[B]].coerce, [4 x <4 x half>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float16x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float16x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast half* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float16x4x4_t, %struct.float16x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL]], i64 0, i64 0 @@ -16834,7 +16834,7 @@ void test_vst1_f16_x4(float16_t *a, float16x4x4_t b) { // CHECK: store [4 x <2 x float>] [[B]].coerce, [4 x <2 x float>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float32x2x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float32x2x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast float* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float32x2x4_t, %struct.float32x2x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <2 x float>], [4 x <2 x float>]* [[VAL]], i64 0, i64 0 @@ -16870,7 +16870,7 @@ void test_vst1_f32_x4(float32_t *a, float32x2x4_t b) { // CHECK: store [4 x <1 x double>] [[B]].coerce, [4 x <1 x double>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.float64x1x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.float64x1x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast double* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.float64x1x4_t, %struct.float64x1x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <1 x double>], [4 x <1 x double>]* [[VAL]], i64 0, i64 0 @@ -16906,7 +16906,7 @@ void test_vst1_f64_x4(float64_t *a, float64x1x4_t b) { // CHECK: store [4 x <8 x i8>] [[B]].coerce, [4 x <8 x i8>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly8x8x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly8x8x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly8x8x4_t, %struct.poly8x8x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <8 x i8>], [4 x <8 x i8>]* [[VAL]], i64 0, i64 0 // CHECK: [[TMP2:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX]], align 8 @@ -16932,7 +16932,7 @@ void test_vst1_p8_x4(poly8_t *a, poly8x8x4_t b) { // CHECK: store [4 x <4 x i16>] [[B]].coerce, [4 x <4 x i16>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly16x4x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly16x4x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i16* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly16x4x4_t, %struct.poly16x4x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <4 x i16>], [4 x <4 x i16>]* [[VAL]], i64 0, i64 0 @@ -16968,7 +16968,7 @@ void test_vst1_p16_x4(poly16_t *a, poly16x4x4_t b) { // CHECK: store [4 x <1 x i64>] [[B]].coerce, [4 x <1 x i64>]* [[COERCE_DIVE]], align 8 // CHECK: [[TMP0:%.*]] = bitcast %struct.poly64x1x4_t* [[__S1]] to i8* // CHECK: [[TMP1:%.*]] = bitcast %struct.poly64x1x4_t* [[B]] to i8* -// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP0]], i8* [[TMP1]], i64 32, i32 8, i1 false) +// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP0]], i8* align 8 [[TMP1]], i64 32, i1 false) // CHECK: [[TMP2:%.*]] = bitcast i64* %a to i8* // CHECK: [[VAL:%.*]] = getelementptr inbounds %struct.poly64x1x4_t, %struct.poly64x1x4_t* [[__S1]], i32 0, i32 0 // CHECK: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x <1 x i64>], [4 x <1 x i64>]* [[VAL]], i64 0, i64 0 |
