Library/src/util/malloc.c at a426c81a8f33717731d100e9fe84fc43b1a98dd7 · AlphaSparse/Library · GitHub

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
/**
 * @brief implement for malloc utils
 * @author Zhuoqiang Guo <gzq9425@qq.com>
 */

#include "alphasparse/util/malloc.h"

#include <stdlib.h>
#include <stdio.h>
#include <time.h>

#include "alphasparse/util/random.h"
#include "alphasparse/util/thread.h"

#ifdef NUMA
#include <numa.h>
#endif

void *alpha_malloc(size_t bytes) {
#ifdef NUMA
  void *ret = numa_alloc_onnode(bytes, 0);
#else
  void *ret = malloc(bytes);
#endif
  if (ret == NULL) {
    printf("no enough memory space to alloc!!!\n");
    exit(-1);
  }
  return ret;
}

void *alpha_memalign(size_t bytes, size_t alignment) {
#ifdef NUMA
  void *ret = numa_alloc_onnode(bytes, 0);
#else
  void *ret = memalign(alignment, bytes);
#endif
  if (ret == NULL) {
    printf("no enough memory space to alloc!!!");
    exit(-1);
  }
  return ret;
}

void alpha_free(void *point) {
  if (!point)
    free(point);
}

void alpha_clear_cache() {
  ALPHA_INT thread_num = alpha_get_thread_num();
  const size_t L3_used = (thread_num + 23) / 24;
  const size_t size = L3_CACHE_SIZE * 8 * L3_used;
  long long *c = (long long *)alpha_memalign(size, DEFAULT_ALIGNMENT);
#ifdef _OPENMP
#pragma omp parallel for num_threads(thread_num)
#endif
  for (size_t i = 0; i < 15; i++)
    for (size_t j = 0; j < L3_CACHE_SIZE * L3_used; j++) c[j] += i * j;
  alpha_free(c);
}

void alpha_fill_s(float *arr, const float num, const size_t size) {
  for (size_t i = 0; i < size; ++i) arr[i] = num;
}

void alpha_fill_d(double *arr, const double num, const size_t size) {
  for (size_t i = 0; i < size; ++i) arr[i] = num;
}

void alpha_fill_c(ALPHA_Complex8 *arr, const ALPHA_Complex8 num, const size_t size) {
  for (size_t i = 0; i < size; ++i) arr[i] = num;
}

void alpha_fill_z(ALPHA_Complex16 *arr, const ALPHA_Complex16 num,
                const size_t size) {
  for (size_t i = 0; i < size; ++i) arr[i] = num;
}

void alpha_fill_random_int(int *arr, unsigned int seed, const size_t size,
                         int upper) {
  if (seed == 0) seed = time_seed();
  srand(seed);
  for (size_t i = 0; i < size; ++i) arr[i] = random_int(upper);
}
void alpha_fill_random_long(long long *arr, unsigned int seed, const size_t size,
                          long long upper) {
  if (seed == 0) seed = time_seed();
  srand(seed);
  for (size_t i = 0; i < size; ++i) arr[i] = random_long(upper);
}
void alpha_fill_random_s(float *arr, unsigned int seed, const size_t size) {
  if (seed == 0) seed = time_seed();
  srand(seed);
  for (size_t i = 0; i < size; ++i) arr[i] = random_float();
}

void alpha_fill_random_d(double *arr, unsigned int seed, const size_t size) {
  if (seed == 0) seed = time_seed();
  srand(seed);
  for (size_t i = 0; i < size; ++i) arr[i] = random_double();
}

void alpha_fill_random_c(ALPHA_Complex8 *arr, unsigned int seed,
                       const size_t size) {
  alpha_fill_random_s((float *)arr, seed, size * 2);
}

void alpha_fill_random_z(ALPHA_Complex16 *arr, unsigned int seed,
                       const size_t size) {
  alpha_fill_random_d((double *)arr, seed, size * 2);
}

void alpha_parallel_fill_random_s(float *arr, unsigned int seed,
                                const size_t size) {
  if (seed == 0) seed = time_seed();
  srand(seed);
  ALPHA_INT thread_num = alpha_get_thread_num();
#ifdef _OPENMP
#pragma omp parallel for num_threads(thread_num)
#endif
  for (size_t i = 0; i < size; ++i) arr[i] = random_float();
}

void alpha_parallel_fill_random_d(double *arr, unsigned int seed,
                                const size_t size) {
  if (seed == 0) seed = time_seed();
  srand(seed);
  ALPHA_INT thread_num = alpha_get_thread_num();
#ifdef _OPENMP
#pragma omp parallel for num_threads(thread_num)
#endif
  for (size_t i = 0; i < size; ++i) arr[i] = random_double();
}

void alpha_parallel_fill_random_c(ALPHA_Complex8 *arr, unsigned int seed,
                                const size_t size) {
  alpha_fill_random_s((float *)arr, seed, size * 2);
}

void alpha_parallel_fill_random_z(ALPHA_Complex16 *arr, unsigned int seed,
                                const size_t size) {
  alpha_fill_random_d((double *)arr, seed, size * 2);
}

void alpha_parallel_fill_s(float *arr, const float num, const size_t size) {
  ALPHA_INT thread_num = alpha_get_thread_num();
#ifdef _OPENMP
#pragma omp parallel for num_threads(thread_num)
#endif
  for (size_t i = 0; i < size; ++i) arr[i] = num;
}

void alpha_parallel_fill_d(double *arr, const double num, const size_t size) {
  ALPHA_INT thread_num = alpha_get_thread_num();
#ifdef _OPENMP
#pragma omp parallel for num_threads(thread_num)
#endif
  for (size_t i = 0; i < size; ++i) arr[i] = num;
}

void alpha_parallel_fill_c(ALPHA_Complex8 *arr, const ALPHA_Complex8 num,
                         const size_t size) {
  ALPHA_INT thread_num = alpha_get_thread_num();
#ifdef _OPENMP
#pragma omp parallel for num_threads(thread_num)
#endif
  for (size_t i = 0; i < size; ++i) arr[i] = num;
}

void alpha_parallel_fill_z(ALPHA_Complex16 *arr, const ALPHA_Complex16 num,
                         const size_t size) {
  ALPHA_INT thread_num = alpha_get_thread_num();
#ifdef _OPENMP
#pragma omp parallel for num_threads(thread_num)
#endif
  for (size_t i = 0; i < size; ++i) arr[i] = num;
}