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
|
/* { dg-do compile } */
/* { dg-options "-O2 -ftree-vectorize -march=armv8-a+sve -msve-vector-bits=scalable" } */
#include <stdint.h>
#define VEC_PERM(TYPE) \
TYPE __attribute__ ((weak)) \
vec_slp_##TYPE (TYPE *restrict a, int n) \
{ \
for (int i = 0; i < n; ++i) \
{ \
a[i * 8] += 99; \
a[i * 8 + 1] += 11; \
a[i * 8 + 2] += 17; \
a[i * 8 + 3] += 80; \
a[i * 8 + 4] += 63; \
a[i * 8 + 5] += 37; \
a[i * 8 + 6] += 24; \
a[i * 8 + 7] += 81; \
} \
}
#define TEST_ALL(T) \
T (int8_t) \
T (uint8_t) \
T (int16_t) \
T (uint16_t) \
T (int32_t) \
T (uint32_t) \
T (int64_t) \
T (uint64_t) \
T (float) \
T (double)
TEST_ALL (VEC_PERM)
/* 1 for each 8-bit type, 2 for each 16-bit type, 4 for each 32-bit type
and 8 for double. */
/* { dg-final { scan-assembler-times {\tld1rd\tz[0-9]+\.d, } 26 } } */
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, #99\n} 2 } } */
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, #11\n} 2 } } */
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, #17\n} 2 } } */
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, #80\n} 2 } } */
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, #63\n} 2 } } */
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, #37\n} 2 } } */
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, #24\n} 2 } } */
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, #81\n} 2 } } */
/* The 16-bit types need 1 ZIP1 each. The 32-bit types need:
ZIP1 ZIP1 (2 ZIP2s optimized away)
ZIP1 ZIP2
and the 64-bit types need:
ZIP1 ZIP1 ZIP1 ZIP1 (4 ZIP2s optimized away)
ZIP1 ZIP2 ZIP1 ZIP2
ZIP1 ZIP2 ZIP1 ZIP2. */
/* { dg-final { scan-assembler-times {\tzip1\tz[0-9]+\.d, z[0-9]+\.d, z[0-9]+\.d\n} 35 } } */
/* { dg-final { scan-assembler-times {\tzip2\tz[0-9]+\.d, z[0-9]+\.d, z[0-9]+\.d\n} 15 } } */
/* The loop should be fully-masked. The 32-bit types need two loads
and stores each and the 64-bit types need four. */
/* { dg-final { scan-assembler-times {\tld1b\t} 2 } } */
/* { dg-final { scan-assembler-times {\tst1b\t} 2 } } */
/* { dg-final { scan-assembler-times {\tld1h\t} 2 } } */
/* { dg-final { scan-assembler-times {\tst1h\t} 2 } } */
/* { dg-final { scan-assembler-times {\tld1w\t} 6 } } */
/* { dg-final { scan-assembler-times {\tst1w\t} 6 } } */
/* { dg-final { scan-assembler-times {\tld1d\t} 12 } } */
/* { dg-final { scan-assembler-times {\tst1d\t} 12 } } */
/* { dg-final { scan-assembler-times {\twhilelo\tp[0-7]\.b} 4 } } */
/* { dg-final { scan-assembler-times {\twhilelo\tp[0-7]\.h} 4 } } */
/* { dg-final { scan-assembler-times {\twhilelo\tp[0-7]\.s} 12 } } */
/* { dg-final { scan-assembler-times {\twhilelo\tp[0-7]\.d} 24 } } */
/* { dg-final { scan-assembler-not {\tldr} } } */
/* { dg-final { scan-assembler-not {\tstr} } } */
/* { dg-final { scan-assembler-not {\tuqdec[bh]\t} } } */
/* We use UQDECW instead of UQDECD ..., MUL #2. */
/* { dg-final { scan-assembler-times {\tuqdecw\t} 6 } } */
/* { dg-final { scan-assembler-times {\tuqdecd\t} 6 } } */
|
