1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
|
// RUN: %clang_cc1 %std_cxx11-14 %s -O0 -triple=x86_64-apple-darwin -target-feature +avx2 -fmax-type-align=16 -emit-llvm -o - -Werror | FileCheck %s --check-prefixes=CHECK,PRE17
// RUN: %clang_cc1 %std_cxx17- %s -O0 -triple=x86_64-apple-darwin -target-feature +avx2 -fmax-type-align=16 -emit-llvm -o - -Werror | FileCheck %s --check-prefixes=CHECK,CXX17
typedef float AVX2Float __attribute__((__vector_size__(32)));
#if __cplusplus < 202002L
volatile
#endif
float TestAlign(void)
{
volatile AVX2Float *p = new AVX2Float;
*p = *p;
AVX2Float r = *p;
return r[0];
}
// CHECK: [[R:%.*]] = alloca <8 x float>, align 32
// PRE17-NEXT: [[CALL:%.*]] = call noalias noundef nonnull ptr @_Znwm(i64 noundef 32)
// CXX17-NEXT: [[CALL:%.*]] = call noalias noundef nonnull align 32 ptr @_ZnwmSt11align_val_t(i64 noundef 32, i64 noundef 32)
// CHECK-NEXT: store ptr [[CALL]], ptr [[P:%.*]], align 8
// CHECK-NEXT: [[ONE:%.*]] = load ptr, ptr [[P]], align 8
// CHECK-NEXT: [[TWO:%.*]] = load volatile <8 x float>, ptr [[ONE]], align 16
// CHECK-NEXT: [[THREE:%.*]] = load ptr, ptr [[P]], align 8
// CHECK-NEXT: store volatile <8 x float> [[TWO]], ptr [[THREE]], align 16
// CHECK-NEXT: [[FOUR:%.*]] = load ptr, ptr [[P]], align 8
// CHECK-NEXT: [[FIVE:%.*]] = load volatile <8 x float>, ptr [[FOUR]], align 16
// CHECK-NEXT: store <8 x float> [[FIVE]], ptr [[R]], align 32
// CHECK-NEXT: [[SIX:%.*]] = load <8 x float>, ptr [[R]], align 32
// CHECK-NEXT: [[VECEXT:%.*]] = extractelement <8 x float> [[SIX]], i32 0
// CHECK-NEXT: ret float [[VECEXT]]
typedef float AVX2Float_Explicitly_aligned __attribute__((__vector_size__(32))) __attribute__((aligned (32)));
typedef AVX2Float_Explicitly_aligned AVX2Float_indirect;
typedef AVX2Float_indirect AVX2Float_use_existing_align;
#if __cplusplus < 202002L
volatile
#endif
float TestAlign2(void)
{
volatile AVX2Float_use_existing_align *p = new AVX2Float_use_existing_align;
*p = *p;
AVX2Float_use_existing_align r = *p;
return r[0];
}
// CHECK: [[R:%.*]] = alloca <8 x float>, align 32
// PRE17-NEXT: [[CALL:%.*]] = call noalias noundef nonnull ptr @_Znwm(i64 noundef 32)
// CXX17-NEXT: [[CALL:%.*]] = call noalias noundef nonnull align 32 ptr @_ZnwmSt11align_val_t(i64 noundef 32, i64 noundef 32)
// CHECK-NEXT: store ptr [[CALL]], ptr [[P:%.*]], align 8
// CHECK-NEXT: [[ONE:%.*]] = load ptr, ptr [[P]], align 8
// CHECK-NEXT: [[TWO:%.*]] = load volatile <8 x float>, ptr [[ONE]], align 32
// CHECK-NEXT: [[THREE:%.*]] = load ptr, ptr [[P]], align 8
// CHECK-NEXT: store volatile <8 x float> [[TWO]], ptr [[THREE]], align 32
// CHECK-NEXT: [[FOUR:%.*]] = load ptr, ptr [[P]], align 8
// CHECK-NEXT: [[FIVE:%.*]] = load volatile <8 x float>, ptr [[FOUR]], align 32
// CHECK-NEXT: store <8 x float> [[FIVE]], ptr [[R]], align 32
// CHECK-NEXT: [[SIX:%.*]] = load <8 x float>, ptr [[R]], align 32
// CHECK-NEXT: [[VECEXT:%.*]] = extractelement <8 x float> [[SIX]], i32 0
// CHECK-NEXT: ret float [[VECEXT]]
|
