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
|
// 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/flops.hh"
#include "print_matrix.hh"
#include "error.hh"
#include "lapacke_wrappers.hh"
#include <vector>
// -----------------------------------------------------------------------------
template< typename scalar_t >
void test_ungqr_work( Params& params, bool run )
{
using real_t = blas::real_type< scalar_t >;
// get & mark input values
int64_t m = params.dim.m();
int64_t n = params.dim.n();
int64_t k = params.dim.k();
int64_t align = params.align();
params.matrix.mark();
real_t eps = std::numeric_limits< real_t >::epsilon();
real_t tol = params.tol() * eps;
// mark non-standard output values
params.ortho();
params.gflops();
params.ref_time();
params.ref_gflops();
params.msg();
if (! run)
return;
// skip invalid sizes
if (! (n <= m && k <= n)) {
params.msg() = "skipping: requires n <= m and k <= n";
return;
}
// ---------- setup
int64_t lda = roundup( blas::max( 1, m ), align );
size_t size_A = (size_t) lda * n;
size_t size_tau = (size_t) (k);
std::vector< scalar_t > A_tst( size_A );
std::vector< scalar_t > A_ref( size_A );
std::vector< scalar_t > A_factored( size_A );
std::vector< scalar_t > tau( size_tau );
lapack::generate_matrix( params.matrix, m, n, &A_tst[0], lda );
int64_t idist = 1;
int64_t iseed[4] = { 0, 1, 2, 3 };
lapack::larnv( idist, iseed, tau.size(), &tau[0] );
// save matrix
A_ref = A_tst;
// ---------- factor matrix
int64_t info_qrf = lapack::geqrf( m, n, &A_tst[0], lda, &tau[0] );
if (info_qrf != 0) {
fprintf( stderr, "lapack::unqrf returned error %lld\n", llong( info_qrf ) );
}
// save matrix
A_factored = A_tst;
// // ---------- run test
testsweeper::flush_cache( params.cache() );
double time = testsweeper::get_wtime();
int64_t info_tst = lapack::ungqr( m, n, k, &A_tst[0], lda, &tau[0] );
time = testsweeper::get_wtime() - time;
if (info_tst != 0) {
fprintf( stderr, "lapack::ungqr returned error %lld\n", llong( info_tst ) );
}
params.time() = time;
double gflop = lapack::Gflop< scalar_t >::ungqr( m, n, k );
params.gflops() = gflop / time;
if (params.check() == 'y') {
// ---------- check error
// comparing to ref. solution doesn't work
// Following lapack/TESTING/LIN/zqrt02.f
// Note (0 <= n <= m) (0 <= k <= n).
int64_t ldq = blas::max( m, n );
int64_t ldr = blas::max( m, n );
std::vector< scalar_t > Q( ldq * n );
std::vector< scalar_t > R( ldr * n );
// Copy the first k columns of the factorization to the array Q
real_t rogue = -10000000000; // -1D+10
lapack::laset( lapack::MatrixType::General, m, n, rogue, rogue, &Q[0], ldq );
lapack::lacpy( lapack::MatrixType::Lower, m-1, k, &A_factored[1], lda, &Q[1], ldq );
// Generate the first n columns of the matrix Q
int64_t info_ungqr = lapack::ungqr( m, n, k, &Q[0], ldq, &tau[0] );
if (info_ungqr != 0) {
fprintf( stderr, "lapack::ungqr returned error %lld\n", llong( info_ungqr ) );
}
// Copy R(1:n,1:k)
lapack::laset( lapack::MatrixType::General, n, k, 0.0, 0.0, &R[0], ldr );
lapack::lacpy( lapack::MatrixType::Upper, n, k, &A_factored[0], lda, &R[0], ldr );
// Compute R - Q'*A
blas::gemm( blas::Layout::ColMajor,
blas::Op::ConjTrans, blas::Op::NoTrans, n, k, m,
-1.0, &Q[0], ldq, &A_ref[0], lda, 1.0, &R[0], ldr );
// Compute norm( R - Q'*A ) / ( M * norm(A) * EPS ) .
real_t Anorm = lapack::lange( lapack::Norm::One, m, k, &A_ref[0], lda );
real_t resid1 = lapack::lange( lapack::Norm::One, n, k, &R[0], ldr );
real_t error1 = 0;
if (Anorm > 0)
error1 = resid1 / ( m * Anorm );
// Compute I - Q'*Q
lapack::laset( lapack::MatrixType::General, n, n, 0.0, 1.0, &R[0], ldr );
blas::herk( blas::Layout::ColMajor, blas::Uplo::Upper, blas::Op::ConjTrans,
n, m, -1.0, &Q[0], ldq, 1.0, &R[0], ldr );
// Compute norm( I - Q'*Q ) / ( M * EPS ) .
real_t resid2 = lapack::lansy( lapack::Norm::One, lapack::Uplo::Upper, n, &R[0], ldr );
real_t error2 = ( resid2 / m );
params.error() = error1;
params.ortho() = error2;
params.okay() = (error1 < tol) && (error2 < tol);
}
if (params.ref() == 'y') {
// ---------- run reference
testsweeper::flush_cache( params.cache() );
time = testsweeper::get_wtime();
int64_t info_ref = LAPACKE_ungqr( m, n, k, &A_ref[0], lda, &tau[0] );
time = testsweeper::get_wtime() - time;
if (info_ref != 0) {
fprintf( stderr, "LAPACKE_ungqr returned error %lld\n", llong( info_ref ) );
}
params.ref_time() = time;
params.ref_gflops() = gflop / time;
}
}
// -----------------------------------------------------------------------------
void test_ungqr( Params& params, bool run )
{
switch (params.datatype()) {
case testsweeper::DataType::Single:
test_ungqr_work< float >( params, run );
break;
case testsweeper::DataType::Double:
test_ungqr_work< double >( params, run );
break;
case testsweeper::DataType::SingleComplex:
test_ungqr_work< std::complex<float> >( params, run );
break;
case testsweeper::DataType::DoubleComplex:
test_ungqr_work< std::complex<double> >( params, run );
break;
default:
throw std::runtime_error( "unknown datatype" );
break;
}
}
|
