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
|
/*
* Copyright (c) 2018 Cisco and/or its affiliates.
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef included_vector_avx2_h
#define included_vector_avx2_h
#include <vppinfra/clib.h>
#include <x86intrin.h>
#define foreach_avx2_vec256i \
_(i,8,32,epi8) _(i,16,16,epi16) _(i,32,8,epi32) _(i,64,4,epi64x)
#define foreach_avx2_vec256u \
_(u,8,32,epi8) _(u,16,16,epi16) _(u,32,8,epi32) _(u,64,4,epi64x)
#define foreach_avx2_vec256f \
_(f,32,8,ps) _(f,64,4,pd)
/* splat, load_unaligned, store_unaligned, is_all_zero, is_all_equal */
#define _(t, s, c, i) \
static_always_inline t##s##x##c \
t##s##x##c##_splat (t##s x) \
{ return (t##s##x##c) _mm256_set1_##i (x); } \
\
static_always_inline t##s##x##c \
t##s##x##c##_load_unaligned (void *p) \
{ return (t##s##x##c) _mm256_loadu_si256 (p); } \
\
static_always_inline void \
t##s##x##c##_store_unaligned (t##s##x##c v, void *p) \
{ _mm256_storeu_si256 ((__m256i *) p, (__m256i) v); } \
\
static_always_inline int \
t##s##x##c##_is_all_zero (t##s##x##c x) \
{ return _mm256_testz_si256 ((__m256i) x, (__m256i) x); } \
\
static_always_inline int \
t##s##x##c##_is_all_equal (t##s##x##c v, t##s x) \
{ return t##s##x##c##_is_all_zero (v != t##s##x##c##_splat (x)); }; \
\
foreach_avx2_vec256i foreach_avx2_vec256u
#undef _
always_inline u32x8
u32x8_permute (u32x8 v, u32x8 idx)
{
return (u32x8) _mm256_permutevar8x32_epi32 ((__m256i) v, (__m256i) idx);
}
always_inline u32x4
u32x8_extract_lo (u32x8 v)
{
return (u32x4) _mm256_extracti128_si256 ((__m256i) v, 0);
}
always_inline u32x4
u32x8_extract_hi (u32x8 v)
{
return (u32x4) _mm256_extracti128_si256 ((__m256i) v, 1);
}
always_inline u32x8
u32x8_insert_lo (u32x8 v1, u32x4 v2)
{
return (u32x8) _mm256_inserti128_si256 ((__m256i) v1, (__m128i) v2, 0);
}
always_inline u32x8
u32x8_insert_hi (u32x8 v1, u32x4 v2)
{
return (u32x8) _mm256_inserti128_si256 ((__m256i) v1, (__m128i) v2, 1);
}
/* _extend_to_ */
/* *INDENT-OFF* */
#define _(f,t,i) \
static_always_inline t \
f##_extend_to_##t (f x) \
{ return (t) _mm256_cvt##i ((__m128i) x); }
_(u16x8, u32x8, epu16_epi32)
_(u16x8, u64x4, epu16_epi64)
_(u32x4, u64x4, epu32_epi64)
_(u8x16, u16x16, epu8_epi64)
_(u8x16, u32x8, epu8_epi32)
_(u8x16, u64x4, epu8_epi64)
_(i16x8, i32x8, epi16_epi32)
_(i16x8, i64x4, epi16_epi64)
_(i32x4, i64x4, epi32_epi64)
_(i8x16, i16x16, epi8_epi64)
_(i8x16, i32x8, epi8_epi32)
_(i8x16, i64x4, epi8_epi64)
#undef _
/* *INDENT-ON* */
#endif /* included_vector_avx2_h */
/*
* fd.io coding-style-patch-verification: ON
*
* Local Variables:
* eval: (c-set-style "gnu")
* End:
*/
|
