MCKL
Monte Carlo Kernel Library
include
mckl
random
internal
philox_avx2_32_common.hpp
Go to the documentation of this file.
1
//============================================================================
2
// MCKL/include/mckl/random/internal/philox_avx2_32_common.hpp
3
//----------------------------------------------------------------------------
4
// MCKL: Monte Carlo Kernel Library
5
//----------------------------------------------------------------------------
6
// Copyright (c) 2013-2018, Yan Zhou
7
// All rights reserved.
8
//
9
// Redistribution and use in source and binary forms, with or without
10
// modification, are permitted provided that the following conditions are met:
11
//
12
// Redistributions of source code must retain the above copyright notice,
13
// this list of conditions and the following disclaimer.
14
//
15
// Redistributions in binary form must reproduce the above copyright notice,
16
// this list of conditions and the following disclaimer in the documentation
17
// and/or other materials provided with the distribution.
18
//
19
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS AS IS
20
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22
// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
23
// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24
// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25
// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27
// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28
// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29
// POSSIBILITY OF SUCH DAMAGE.
30
//============================================================================
31
32
#ifndef MCKL_RANDOM_INTERNAL_PHILOX_AVX2_32_COMMON_HPP
33
#define MCKL_RANDOM_INTERNAL_PHILOX_AVX2_32_COMMON_HPP
34
35
#define MCKL_RANDOM_INTERNAL_PHILOX_AVX2_32_RBOX(K, N, imm8) \
36
{ \
37
constexpr int m0 = \
38
static_cast<int>(Constants::multiplier::value[0 % (K / 2)]); \
39
constexpr int m1 = \
40
static_cast<int>(Constants::multiplier::value[1 % (K / 2)]); \
41
constexpr int m2 = \
42
static_cast<int>(Constants::multiplier::value[2 % (K / 2)]); \
43
constexpr int m3 = \
44
static_cast<int>(Constants::multiplier::value[3 % (K / 2)]); \
45
const __m256i ymmm = _mm256_set_epi32(0, m3, 0, m2, 0, m1, 0, m0); \
46
const __m256i ymm8 = _mm256_mul_epu32(ymm0, ymmm); \
47
const __m256i ymm9 = _mm256_mul_epu32(ymm1, ymmm); \
48
const __m256i ymmA = _mm256_mul_epu32(ymm2, ymmm); \
49
const __m256i ymmB = _mm256_mul_epu32(ymm3, ymmm); \
50
const __m256i ymmC = _mm256_mul_epu32(ymm4, ymmm); \
51
const __m256i ymmD = _mm256_mul_epu32(ymm5, ymmm); \
52
const __m256i ymmE = _mm256_mul_epu32(ymm6, ymmm); \
53
const __m256i ymmF = _mm256_mul_epu32(ymm7, ymmm); \
54
\
55
constexpr int ma = static_cast<int>(0xFFFFFFFF); \
56
const __m256i ymma = _mm256_set_epi32(ma, 0, ma, 0, ma, 0, ma, 0); \
57
ymm0 = _mm256_and_si256(ymm0, ymma); \
58
ymm1 = _mm256_and_si256(ymm1, ymma); \
59
ymm2 = _mm256_and_si256(ymm2, ymma); \
60
ymm3 = _mm256_and_si256(ymm3, ymma); \
61
ymm4 = _mm256_and_si256(ymm4, ymma); \
62
ymm5 = _mm256_and_si256(ymm5, ymma); \
63
ymm6 = _mm256_and_si256(ymm6, ymma); \
64
ymm7 = _mm256_and_si256(ymm7, ymma); \
65
\
66
constexpr int w0 = \
67
static_cast<int>(Constants::weyl::value[0 % (K / 2)] * N); \
68
constexpr int w1 = \
69
static_cast<int>(Constants::weyl::value[1 % (K / 2)] * N); \
70
constexpr int w2 = \
71
static_cast<int>(Constants::weyl::value[2 % (K / 2)] * N); \
72
constexpr int w3 = \
73
static_cast<int>(Constants::weyl::value[3 % (K / 2)] * N); \
74
const __m256i ymmw = _mm256_set_epi32(w3, 0, w2, 0, w1, 0, w0, 0); \
75
const __m256i ymmk = _mm256_add_epi32(ymmk0, ymmw); \
76
ymm0 = _mm256_xor_si256(ymm0, ymmk); \
77
ymm1 = _mm256_xor_si256(ymm1, ymmk); \
78
ymm2 = _mm256_xor_si256(ymm2, ymmk); \
79
ymm3 = _mm256_xor_si256(ymm3, ymmk); \
80
ymm4 = _mm256_xor_si256(ymm4, ymmk); \
81
ymm5 = _mm256_xor_si256(ymm5, ymmk); \
82
ymm6 = _mm256_xor_si256(ymm6, ymmk); \
83
ymm7 = _mm256_xor_si256(ymm7, ymmk); \
84
\
85
ymm0 = _mm256_xor_si256(ymm0, ymm8); \
86
ymm1 = _mm256_xor_si256(ymm1, ymm9); \
87
ymm2 = _mm256_xor_si256(ymm2, ymmA); \
88
ymm3 = _mm256_xor_si256(ymm3, ymmB); \
89
ymm4 = _mm256_xor_si256(ymm4, ymmC); \
90
ymm5 = _mm256_xor_si256(ymm5, ymmD); \
91
ymm6 = _mm256_xor_si256(ymm6, ymmE); \
92
ymm7 = _mm256_xor_si256(ymm7, ymmF); \
93
\
94
ymm0 = _mm256_shuffle_epi32(ymm0, imm8); \
95
ymm1 = _mm256_shuffle_epi32(ymm1, imm8); \
96
ymm2 = _mm256_shuffle_epi32(ymm2, imm8); \
97
ymm3 = _mm256_shuffle_epi32(ymm3, imm8); \
98
ymm4 = _mm256_shuffle_epi32(ymm4, imm8); \
99
ymm5 = _mm256_shuffle_epi32(ymm5, imm8); \
100
ymm6 = _mm256_shuffle_epi32(ymm6, imm8); \
101
ymm7 = _mm256_shuffle_epi32(ymm7, imm8); \
102
}
103
104
#endif // MCKL_RANDOM_INTERNAL_PHILOX_AVX2_32_COMMON_HPP
Generated by
1.8.13
1.8.13