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
|
// Copyright (c) 2017-2023, University of Tennessee. All rights reserved.
// SPDX-License-Identifier: BSD-3-Clause
// This program is free software: you can redistribute it and/or modify it under
// the terms of the BSD 3-Clause license. See the accompanying LICENSE file.
#include "test.hh"
#include "lapack.hh"
#include "lapack/device.hh"
#include "lapack/flops.hh"
#include "print_matrix.hh"
#include "error.hh"
#include "check_heev.hh"
#include "lapacke_wrappers.hh"
#include <vector>
template< typename scalar_t >
void test_heevd_device_work( Params& params, bool run )
{
using lapack::device_info_int;
using real_t = blas::real_type< scalar_t >;
using lapack::Job;
// Constants
const real_t eps = std::numeric_limits< real_t >::epsilon();
// get & mark input values
lapack::Job jobz = params.jobz();
lapack::Uplo uplo = params.uplo();
int64_t n = params.dim.n();
int64_t device = params.device();
int64_t align = params.align();
int64_t verbose = params.verbose();
real_t tol = params.tol() * eps;
params.matrix.mark();
// mark non-standard output values
params.ref_time();
// params.ref_gflops();
// params.gflops();
params.ortho();
params.error2();
params.error2.name( "Lambda" );
if (! run)
return;
if (blas::get_device_count() == 0) {
params.msg() = "skipping: no GPU devices or no GPU support";
return;
}
// ---------- setup
int64_t lda = roundup( blas::max( 1, n ), align );
int64_t ldz = lda; // vectors overwrite matrix A
size_t size_A = (size_t) lda * n;
size_t size_Z = size_A;
size_t size_W = (size_t) n;
std::vector< scalar_t > A( size_A );
std::vector< scalar_t > Z( size_Z ); // eigenvectors
std::vector< real_t > Lambda_tst( n );
std::vector< real_t > Lambda_ref( n );
lapack::generate_matrix( params.matrix, n, n, &A[0], lda );
Z = A;
// Allocate and copy to GPU
lapack::Queue queue( device );
scalar_t* dA_tst = blas::device_malloc< scalar_t >( size_A, queue );
real_t* dW_tst = blas::device_malloc< real_t > ( size_W, queue );
device_info_int* d_info = blas::device_malloc< device_info_int >( 1, queue );
blas::device_copy_matrix( n, n, A.data(), lda, dA_tst, lda, queue );
// Allocate workspace
size_t d_size, h_size;
lapack::heevd_work_size_bytes( jobz, uplo, n, dA_tst, lda, dW_tst,
&d_size, &h_size, queue );
char* d_work = blas::device_malloc< char >( d_size, queue );
std::vector<char> h_work_vector( h_size );
char* h_work = h_work_vector.data();
if (verbose >= 1) {
printf( "\n" );
printf( "A n=%5lld, lda=%5lld\n", llong( n ), llong( lda ) );
}
if (verbose >= 2) {
printf( "A = " ); print_matrix( n, n, &A[0], lda );
}
// ---------- run test
testsweeper::flush_cache( params.cache() );
queue.sync();
double time = testsweeper::get_wtime();
lapack::heevd( jobz, uplo, n, dA_tst, lda, dW_tst, d_work, d_size,
h_work, h_size, d_info, queue );
queue.sync();
time = testsweeper::get_wtime() - time;
params.time() = time;
// double gflop = lapack::Gflop< scalar_t >::heev( jobz, n );
// params.gflops() = gflop / time;
// Copy result back to CPU.
device_info_int info_tst;
blas::device_copy_matrix( n, n, dA_tst, lda, Z.data(), ldz, queue );
blas::device_copy_vector( n, dW_tst, 1, Lambda_tst.data(), 1, queue );
blas::device_memcpy( &info_tst, d_info, 1, queue );
queue.sync();
if (info_tst != 0) {
fprintf( stderr, "lapack::heev returned error %lld\n", llong( info_tst ) );
}
// Cleanup GPU memory
blas::device_free( dA_tst, queue );
blas::device_free( dW_tst, queue );
blas::device_free( d_work, queue );
blas::device_free( d_info, queue );
if (verbose >= 2) {
printf( "Z = " ); print_matrix( n, n, &Z[0], ldz );
printf( "Lambda = " ); print_vector( n, &Lambda_tst[0], 1 );
}
if (params.check() == 'y') {
// ---------- check numerical error
// result[ 0 ] = || A - Z Lambda Z^H || / (n ||A||), if jobz != NoVec.
// result[ 1 ] = || I - Z^H Z || / n, if jobz != NoVec.
// result[ 2 ] = 0 if Lambda is in non-decreasing order, else > 0.
real_t result[ 3 ] = { (real_t) testsweeper::no_data_flag,
(real_t) testsweeper::no_data_flag,
(real_t) testsweeper::no_data_flag };
check_heev( jobz, uplo, n, &A[0], lda,
n, &Lambda_tst[0], &Z[0], ldz, result );
params.error() = result[ 0 ];
params.ortho() = result[ 1 ];
params.error2() = result[ 2 ];
params.okay() = (jobz == Job::NoVec || result[ 0 ] < tol)
&& (jobz == Job::NoVec || result[ 1 ] < tol)
&& result[ 2 ] < tol;
}
if (params.ref() == 'y' || params.check() == 'y') {
// ---------- run reference
testsweeper::flush_cache( params.cache() );
time = testsweeper::get_wtime();
int64_t info_ref = LAPACKE_heev(
to_char( jobz ), to_char( uplo ), n,
&A[0], lda, &Lambda_ref[0] );
time = testsweeper::get_wtime() - time;
if (info_ref != 0) {
fprintf( stderr, "LAPACKE_heev returned error %lld\n", llong( info_ref ) );
}
params.ref_time() = time;
// params.ref_gflops() = gflop / time;
// ---------- check error compared to reference
real_t error = rel_error( Lambda_tst, Lambda_ref );
if (info_tst != info_ref) {
error = 1;
}
params.error2() = error;
params.okay() = params.okay() && (error < tol);
}
}
// -----------------------------------------------------------------------------
void test_heevd_device( Params& params, bool run )
{
switch (params.datatype()) {
case testsweeper::DataType::Single:
test_heevd_device_work< float >( params, run );
break;
case testsweeper::DataType::Double:
test_heevd_device_work< double >( params, run );
break;
case testsweeper::DataType::SingleComplex:
test_heevd_device_work< std::complex<float> >( params, run );
break;
case testsweeper::DataType::DoubleComplex:
test_heevd_device_work< std::complex<double> >( params, run );
break;
default:
throw std::runtime_error( "unknown datatype" );
break;
}
}
|
