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
|
/*
-- MAGMA (version 2.9.0) --
Univ. of Tennessee, Knoxville
Univ. of California, Berkeley
Univ. of Colorado, Denver
@date January 2025
@author Azzam Haidar
@author Tingxing Dong
@generated from testing/testing_zpotrf_batched.cpp, normal z -> s, Wed Jan 22 14:40:45 2025
*/
// includes, system
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <math.h>
// includes, project
#include "flops.h"
#include "magma_v2.h"
#include "magma_lapack.h"
#include "testings.h"
#if defined(_OPENMP)
#include <omp.h>
#endif
#include "../control/magma_threadsetting.h" // internal header
/* ////////////////////////////////////////////////////////////////////////////
-- Testing spotrf_batched
*/
int main( int argc, char** argv)
{
TESTING_CHECK( magma_init() );
magma_print_environment();
real_Double_t gflops, gpu_perf, gpu_time, cpu_perf, cpu_time;
float *h_A, *h_R;
float *d_A;
magma_int_t N, n2, lda, ldda, info;
float c_neg_one = MAGMA_S_NEG_ONE;
magma_int_t ione = 1;
magma_int_t ISEED[4] = {0,0,0,1};
float work[1], error;
int status = 0;
float **d_A_array = NULL;
magma_int_t *dinfo_magma;
magma_int_t *hinfo_magma;
magma_int_t batchCount;
magma_opts opts( MagmaOptsBatched );
opts.parse_opts( argc, argv );
opts.lapack |= opts.check; // check (-c) implies lapack (-l)
batchCount = opts.batchcount;
float tol = opts.tolerance * lapackf77_slamch("E");
magma_queue_t queue = opts.queue;
printf("%% BatchCount N CPU Gflop/s (ms) GPU Gflop/s (ms) ||R_magma - R_lapack||_F / ||R_lapack||_F\n");
printf("%%===================================================================================================\n");
for( int itest = 0; itest < opts.ntest; ++itest ) {
for( int iter = 0; iter < opts.niter; ++iter ) {
N = opts.nsize[itest];
lda = N;
ldda = magma_roundup( N, opts.align ); // multiple of 32 by default
n2 = lda* N * batchCount;
gflops = batchCount * FLOPS_SPOTRF( N ) / 1e9;
TESTING_CHECK( magma_imalloc_cpu( &hinfo_magma, batchCount ));
TESTING_CHECK( magma_smalloc_cpu( &h_A, n2 ));
TESTING_CHECK( magma_smalloc_pinned( &h_R, n2 ));
TESTING_CHECK( magma_smalloc( &d_A, ldda * N * batchCount ));
TESTING_CHECK( magma_imalloc( &dinfo_magma, batchCount ));
TESTING_CHECK( magma_malloc( (void**) &d_A_array, batchCount * sizeof(float*) ));
/* Initialize the matrix */
lapackf77_slarnv( &ione, ISEED, &n2, h_A );
for (int i=0; i < batchCount; i++)
{
magma_smake_hpd( N, h_A + i * lda * N, lda ); // need modification
}
magma_int_t columns = N * batchCount;
lapackf77_slacpy( MagmaFullStr, &N, &(columns), h_A, &lda, h_R, &lda );
magma_ssetmatrix( N, columns, h_A, lda, d_A, ldda, opts.queue );
/* ====================================================================
Performs operation using MAGMA
=================================================================== */
magma_memset( dinfo_magma, 0, batchCount * sizeof(magma_int_t) );
magma_sset_pointer( d_A_array, d_A, ldda, 0, 0, ldda * N, batchCount, queue );
gpu_time = magma_sync_wtime( opts.queue );
info = magma_spotrf_batched( opts.uplo, N, d_A_array, ldda, dinfo_magma, batchCount, queue);
gpu_time = magma_sync_wtime( opts.queue ) - gpu_time;
gpu_perf = gflops / gpu_time;
magma_getvector( batchCount, sizeof(magma_int_t), dinfo_magma, 1, hinfo_magma, 1, opts.queue );
for (int i=0; i < batchCount; i++)
{
if (hinfo_magma[i] != 0 ) {
printf("magma_spotrf_batched matrix %lld returned diag error %lld\n",
(long long) i, (long long) hinfo_magma[i] );
status = -1;
}
}
if (info != 0) {
//printf("magma_spotrf_batched returned argument error %lld: %s.\n", (long long) info, magma_strerror( info ));
status = -1;
}
if (status == -1)
goto cleanup;
/* =====================================================================
Performs operation using LAPACK
=================================================================== */
if ( opts.lapack ) {
cpu_time = magma_wtime();
// #define BATCHED_DISABLE_PARCPU
#if !defined (BATCHED_DISABLE_PARCPU) && defined(_OPENMP)
magma_int_t nthreads = magma_get_lapack_numthreads();
magma_set_lapack_numthreads(1);
magma_set_omp_numthreads(nthreads);
#pragma omp parallel for schedule(dynamic)
#endif
for (magma_int_t s=0; s < batchCount; s++)
{
magma_int_t locinfo;
lapackf77_spotrf( lapack_uplo_const(opts.uplo), &N, h_A + s * lda * N, &lda, &locinfo );
if (locinfo != 0) {
printf("lapackf77_spotrf matrix %lld returned error %lld: %s.\n",
(long long) s, (long long) locinfo, magma_strerror( locinfo ));
}
}
#if !defined (BATCHED_DISABLE_PARCPU) && defined(_OPENMP)
magma_set_lapack_numthreads(nthreads);
#endif
cpu_time = magma_wtime() - cpu_time;
cpu_perf = gflops / cpu_time;
/* =====================================================================
Check the result compared to LAPACK
=================================================================== */
magma_sgetmatrix( N, columns, d_A, ldda, h_R, lda, opts.queue );
magma_int_t NN = lda*N;
const char* uplo = lapack_uplo_const(opts.uplo);
error = 0;
for (int i=0; i < batchCount; i++)
{
float Anorm, err;
blasf77_saxpy(&NN, &c_neg_one, h_A + i * lda*N, &ione, h_R + i * lda*N, &ione);
Anorm = safe_lapackf77_slansy("f", uplo, &N, h_A + i * lda*N, &lda, work);
err = safe_lapackf77_slansy("f", uplo, &N, h_R + i * lda*N, &lda, work)
/ Anorm;
if (std::isnan(err) || std::isinf(err)) {
error = err;
break;
}
error = max( err, error );
}
bool okay = (error < tol);
status += ! okay;
printf("%10lld %5lld %7.2f (%7.2f) %7.2f (%7.2f) %8.2e %s\n",
(long long) batchCount, (long long) N, cpu_perf, cpu_time*1000., gpu_perf, gpu_time*1000.,
error, (okay ? "ok" : "failed"));
}
else {
printf("%10lld %5lld --- ( --- ) %7.2f (%7.2f) ---\n",
(long long) batchCount, (long long) N, gpu_perf, gpu_time*1000. );
}
cleanup:
magma_free_cpu( hinfo_magma );
magma_free_cpu( h_A );
magma_free_pinned( h_R );
magma_free( d_A );
magma_free( d_A_array );
magma_free( dinfo_magma );
if (status == -1)
break;
fflush( stdout );
}
if (status == -1)
break;
if ( opts.niter > 1 ) {
printf( "\n" );
}
}
opts.cleanup();
TESTING_CHECK( magma_finalize() );
return status;
}
|
