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
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
|
/* { dg-do compile } */
/* { dg-require-effective-target arm_neon_ok } */
/* { dg-require-effective-target arm_little_endian } */
/* { dg-options "-O2" } */
/* { dg-add-options arm_neon } */
#include <arm_neon.h>
uint8x8_t
tst_vext_u8 (uint8x8_t __a, uint8x8_t __b)
{
uint8x8_t __mask1 = {2, 3, 4, 5, 6, 7, 8, 9};
return __builtin_shuffle ( __a, __b, __mask1) ;
}
uint8x8_t
tst_vext_u8_rotate (uint8x8_t __a)
{
uint8x8_t __mask1 = {2, 3, 4, 5, 6, 7, 0, 1};
return __builtin_shuffle ( __a, __mask1) ;
}
uint16x4_t
tst_vext_u16 (uint16x4_t __a, uint16x4_t __b)
{
uint16x4_t __mask1 = {2, 3, 4, 5};
return __builtin_shuffle ( __a, __b, __mask1) ;
}
uint16x4_t
tst_vext_u16_rotate (uint16x4_t __a)
{
uint16x4_t __mask1 = {2, 3, 0, 1};
return __builtin_shuffle ( __a, __mask1) ;
}
uint32x2_t
tst_vext_u32 (uint32x2_t __a, uint32x2_t __b)
{
uint32x2_t __mask1 = {1, 2};
return __builtin_shuffle ( __a, __b, __mask1) ;
}
/* This one is mapped into vrev64.32. */
uint32x2_t
tst_vext_u32_rotate (uint32x2_t __a)
{
uint32x2_t __mask1 = {1, 0};
return __builtin_shuffle ( __a, __mask1) ;
}
uint8x16_t
tst_vextq_u8 (uint8x16_t __a, uint8x16_t __b)
{
uint8x16_t __mask1 = {4, 5, 6, 7, 8, 9, 10, 11,
12, 13, 14, 15, 16, 17, 18, 19};
return __builtin_shuffle ( __a, __b, __mask1) ;
}
uint8x16_t
tst_vextq_u8_rotate (uint8x16_t __a)
{
uint8x16_t __mask1 = {4, 5, 6, 7, 8, 9, 10, 11,
12, 13, 14, 15, 0, 1, 2, 3};
return __builtin_shuffle ( __a, __mask1) ;
}
uint16x8_t
tst_vextq_u16 (uint16x8_t __a, uint16x8_t __b)
{
uint16x8_t __mask1 = {2, 3, 4, 5, 6, 7, 8, 9};
return __builtin_shuffle ( __a, __b, __mask1) ;
}
uint16x8_t
tst_vextq_u16_rotate (uint16x8_t __a)
{
uint16x8_t __mask1 = {2, 3, 4, 5, 6, 7, 0, 1};
return __builtin_shuffle ( __a, __mask1) ;
}
uint32x4_t
tst_vextq_u32 (uint32x4_t __a, uint32x4_t __b)
{
uint32x4_t __mask1 = {1, 2, 3, 4};
return __builtin_shuffle ( __a, __b, __mask1) ;
}
uint32x4_t
tst_vextq_u32_rotate (uint32x4_t __a)
{
uint32x4_t __mask1 = {1, 2, 3, 0};
return __builtin_shuffle ( __a, __mask1) ;
}
uint64x2_t
tst_vextq_u64 (uint64x2_t __a, uint64x2_t __b)
{
uint64x2_t __mask1 = {1, 2};
return __builtin_shuffle ( __a, __b, __mask1) ;
}
uint64x2_t
tst_vextq_u64_rotate (uint64x2_t __a)
{
uint64x2_t __mask1 = {1, 0};
return __builtin_shuffle ( __a, __mask1) ;
}
/* { dg-final {scan-assembler-times "vext\.8\\t" 4} } */
/* { dg-final {scan-assembler-times "vext\.16\\t" 4} } */
/* { dg-final {scan-assembler-times "vext\.32\\t" 3} } */
/* { dg-final {scan-assembler-times "vrev64\.32\\t" 1} } */
/* { dg-final {scan-assembler-times "vext\.64\\t" 2} } */
|
