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
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241 | /*************************************************************************
*
* Project
* _____ _____ __ __ _____
* / ____| __ \| \/ | __ \
* ___ _ __ ___ _ __ | | __| |__) | \ / | |__) |
* / _ \| '_ \ / _ \ '_ \| | |_ | ___/| |\/| | ___/
*| (_) | |_) | __/ | | | |__| | | | | | | |
* \___/| .__/ \___|_| |_|\_____|_| |_| |_|_|
* | |
* |_|
*
* Copyright (C) Akiel Aries, <akiel@akiel.org>, et al.
*
* This software is licensed as described in the file LICENSE, which
* you should have received as part of this distribution. The terms
* among other details are referenced in the official documentation
* seen here : https://akielaries.github.io/openGPMP/ along with
* important files seen in this project.
*
* You may opt to use, copy, modify, merge, publish, distribute
* and/or sell copies of the Software, and permit persons to whom
* the Software is furnished to do so, under the terms of the
* LICENSE file. As this is an Open Source effort, all implementations
* must be of the same methodology.
*
*
*
* This software is distributed on an AS IS basis, WITHOUT
* WARRANTY OF ANY KIND, either express or implied.
*
************************************************************************/
#include <cassert>
#include <cstddef>
#include <cstdint>
#include <iostream>
#include <openGPMP/linalg/mtx.hpp>
#include <vector>
#if defined(__x86_64__) || defined(__amd64__) || defined(__amd64)
/************************************************************************
*
* Matrix Operations for AVX ISA
*
************************************************************************/
#if defined(__AVX2__)
// AVX family intrinsics
#include <immintrin.h>
/************************************************************************
*
* Matrix Operations on vector<vector>
*
************************************************************************/
// matrix addition using Intel intrinsic, accepts integer types
void gpmp::linalg::Mtx::mtx_add(const std::vector<std::vector<int>> &A,
const std::vector<std::vector<int>> &B,
std::vector<std::vector<int>> &C) {
const int rows = A.size();
const int cols = A[0].size();
if (rows > 16) {
for (int i = 0; i < rows; ++i) {
int j = 0;
// requires at least size 8x8 size matrices
for (; j < cols - 7; j += 8) {
// load 8 elements from A, B, and C matrices using SIMD
__m256i a = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&A[i][j]));
__m256i b = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&B[i][j]));
__m256i c = _mm256_loadu_si256(<--- c is initialized
reinterpret_cast<const __m256i *>(&C[i][j]));
// perform vectorized addition
c = _mm256_add_epi32(a, b);<--- c is overwritten
// store the result back to the C matrix
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&C[i][j]), c);
}
// handle the remaining elements that are not multiples of 8
for (; j < cols; ++j) {
C[i][j] = A[i][j] + B[i][j];
}
}
}
// use standard matrix addition
else {
std_mtx_add(A, B, C);
}
}
// matrix subtraction using Intel intrinsics, accepts integer types
void gpmp::linalg::Mtx::mtx_sub(const std::vector<std::vector<int>> &A,
const std::vector<std::vector<int>> &B,
std::vector<std::vector<int>> &C) {
const int rows = A.size();
const int cols = A[0].size();
for (int i = 0; i < rows; ++i) {
int j = 0;
// requires at least size 8x8 size matrices
for (; j < cols - 7; j += 8) {
// load 8 elements from A, B, and C matrices using SIMD
__m256i a =
_mm256_loadu_si256(reinterpret_cast<const __m256i *>(&A[i][j]));
__m256i b =
_mm256_loadu_si256(reinterpret_cast<const __m256i *>(&B[i][j]));
__m256i c =<--- c is initialized
_mm256_loadu_si256(reinterpret_cast<const __m256i *>(&C[i][j]));
// perform vectorized subtraction
c = _mm256_sub_epi32(a, b);<--- c is overwritten
// store the result back to the C matrix
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&C[i][j]), c);
}
// handle the remaining elements that are not multiples of 8
for (; j < cols; ++j) {
C[i][j] = A[i][j] - B[i][j];
}
}
}
// matrix multiplication using Intel intrinsics, accepts integer types
void gpmp::linalg::Mtx::mtx_mult(const std::vector<std::vector<int>> &A,
const std::vector<std::vector<int>> &B,
std::vector<std::vector<int>> &C) {
const int rows_a = A.size();
const int cols_a = A[0].size();
const int cols_b = B[0].size();
for (int i = 0; i < rows_a; ++i) {
for (int j = 0; j < cols_b; j += 8) {
// initialize a vector of zeros for the result
__m256i c = _mm256_setzero_si256();
for (int k = 0; k < cols_a; ++k) {
// load 8 elements from matrices A and B using SIMD
__m256i a = _mm256_set1_epi32(A[i][k]);
__m256i b = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&B[k][j]));
// perform vectorized multiplication
__m256i prod = _mm256_mullo_epi32(a, b);
// perform vectorized addition
c = _mm256_add_epi32(c, prod);
}
// store the result back to the C matrix
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&C[i][j]), c);
}
// handle the remaining elements that are not multiples of 8
for (int j = cols_b - cols_b % 8; j < cols_b; ++j) {
int sum = 0;
for (int k = 0; k < cols_a; ++k) {
sum += A[i][k] * B[k][j];
}
C[i][j] = sum;
}
}
}
// transpose matrices of type int using Intel intrinsics
void gpmp::linalg::Mtx::mtx_tpose(std::vector<std::vector<int>> &matrix) {
const int rows = matrix.size();
const int cols = matrix[0].size();
for (int i = 0; i < rows; i += 8) {
for (int j = i; j < cols; j += 8) {
__m256i row1 = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&matrix[i][j]));
__m256i row2 = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&matrix[i + 1][j]));
__m256i row3 = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&matrix[i + 2][j]));
__m256i row4 = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&matrix[i + 3][j]));
__m256i row5 = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&matrix[i + 4][j]));
__m256i row6 = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&matrix[i + 5][j]));
__m256i row7 = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&matrix[i + 6][j]));
__m256i row8 = _mm256_loadu_si256(
reinterpret_cast<const __m256i *>(&matrix[i + 7][j]));
__m256i tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8;
// transpose 8x8 submatrix
tmp1 = _mm256_unpacklo_epi32(row1, row2);
tmp2 = _mm256_unpacklo_epi32(row3, row4);
tmp3 = _mm256_unpacklo_epi32(row5, row6);
tmp4 = _mm256_unpacklo_epi32(row7, row8);
tmp5 = _mm256_unpackhi_epi32(row1, row2);
tmp6 = _mm256_unpackhi_epi32(row3, row4);
tmp7 = _mm256_unpackhi_epi32(row5, row6);
tmp8 = _mm256_unpackhi_epi32(row7, row8);
row1 = _mm256_unpacklo_epi64(tmp1, tmp2);
row2 = _mm256_unpacklo_epi64(tmp3, tmp4);
row3 = _mm256_unpackhi_epi64(tmp1, tmp2);
row4 = _mm256_unpackhi_epi64(tmp3, tmp4);
row5 = _mm256_unpacklo_epi64(tmp5, tmp6);
row6 = _mm256_unpacklo_epi64(tmp7, tmp8);
row7 = _mm256_unpackhi_epi64(tmp5, tmp6);
row8 = _mm256_unpackhi_epi64(tmp7, tmp8);
// store the transposed 8x8 submatrix back to the matrix
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&matrix[i][j]),
row1);
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&matrix[i + 1][j]),
row2);
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&matrix[i + 2][j]),
row3);
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&matrix[i + 3][j]),
row4);
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&matrix[i + 4][j]),
row5);
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&matrix[i + 5][j]),
row6);
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&matrix[i + 6][j]),
row7);
_mm256_storeu_si256(reinterpret_cast<__m256i *>(&matrix[i + 7][j]),
row8);
}
}
}
#endif
#endif
|