1 /*
2 * Copyright (c) 2009, Michael Lehn
3 *
4 * All rights reserved.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 *
10 * 1) Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2) Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in
14 * the documentation and/or other materials provided with the
15 * distribution.
16 * 3) Neither the name of the FLENS development group nor the names of
17 * its contributors may be used to endorse or promote products derived
18 * from this software without specific prior written permission.
19 *
20 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
23 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
24 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
25 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
26 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
27 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
28 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
29 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
30 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31 */
32
33 #ifndef CXXBLAS_LEVEL1_NRM2_TCC
34 #define CXXBLAS_LEVEL1_NRM2_TCC 1
35
36 #include <cxxblas/aux/complex.h>
37
38 namespace cxxblas {
39
40 template <typename IndexType, typename X, typename T>
41 void
42 nrm2_generic(IndexType n, const X *x, IndexType incX, T &norm)
43 {
44 CXXBLAS_DEBUG_OUT("nrm2_generic");
45
46 using std::abs;
47 using std::pow;
48 using std::sqrt;
49
50 if (n<1) {
51 norm = T(0);
52 } else if (n==1) {
53 norm = abs(*x);
54 } else {
55 T scale = 0;
56 T ssq = 1;
57 // The following loop is equivalent to this call to the LAPACK
58 // auxiliary routine:
59 // CALL DLASSQ( N, X, INCX, SCALE, SSQ )
60 //
61 for (IndexType i=0, iX=0; i<n; ++i, iX+=incX) {
62 if (x[iX]!=T(0)) {
63 T absXi = abs(x[iX]);
64 if (scale<absXi) {
65 ssq = T(1) + ssq * pow(scale/absXi, 2);
66 scale = absXi;
67 } else {
68 ssq += pow(absXi/scale, 2);
69 }
70 }
71 }
72 norm = scale*sqrt(ssq);
73 }
74 }
75
76 template <typename IndexType, typename X, typename T>
77 void
78 nrm2(IndexType n, const X *x, IndexType incX, T &norm)
79 {
80 if (incX<0) {
81 x -= incX*(n-1);
82 }
83 nrm2_generic(n, x, incX, norm);
84 }
85
86 #ifdef HAVE_CBLAS
87
88 // snrm2
89 template <typename IndexType>
90 typename If<IndexType>::isBlasCompatibleInteger
91 nrm2(IndexType n, const float *x, IndexType incX, float &norm)
92 {
93 CXXBLAS_DEBUG_OUT("[" BLAS_IMPL "] cblas_snrm2");
94
95 norm = cblas_snrm2(n, x, incX);
96 }
97
98 // dnrm2
99 template <typename IndexType>
100 typename If<IndexType>::isBlasCompatibleInteger
101 nrm2(IndexType n, const double *x, IndexType incX, double &norm)
102 {
103 CXXBLAS_DEBUG_OUT("[" BLAS_IMPL "] cblas_dnrm2");
104
105 norm = cblas_dnrm2(n, x, incX);
106 }
107
108 // scnrm2
109 template <typename IndexType>
110 typename If<IndexType>::isBlasCompatibleInteger
111 nrm2(IndexType n, const ComplexFloat *x, IndexType incX, float &norm)
112 {
113 CXXBLAS_DEBUG_OUT("[" BLAS_IMPL "] cblas_scnrm2");
114
115 norm = cblas_scnrm2(n, x, incX);
116 }
117
118 // dznrm2
119 template <typename IndexType>
120 typename If<IndexType>::isBlasCompatibleInteger
121 nrm2(IndexType n, const ComplexDouble *x, IndexType incX, double &norm)
122 {
123 CXXBLAS_DEBUG_OUT("[" BLAS_IMPL "] cblas_dznrm2");
124
125 norm = cblas_dznrm2(n, x, incX);
126 }
127
128 #endif // HAVE_CBLAS
129
130 } // namespace cxxblas
131
132 #endif // CXXBLAS_LEVEL1_NRM2_TCC
2 * Copyright (c) 2009, Michael Lehn
3 *
4 * All rights reserved.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 *
10 * 1) Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2) Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in
14 * the documentation and/or other materials provided with the
15 * distribution.
16 * 3) Neither the name of the FLENS development group nor the names of
17 * its contributors may be used to endorse or promote products derived
18 * from this software without specific prior written permission.
19 *
20 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
23 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
24 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
25 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
26 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
27 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
28 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
29 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
30 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31 */
32
33 #ifndef CXXBLAS_LEVEL1_NRM2_TCC
34 #define CXXBLAS_LEVEL1_NRM2_TCC 1
35
36 #include <cxxblas/aux/complex.h>
37
38 namespace cxxblas {
39
40 template <typename IndexType, typename X, typename T>
41 void
42 nrm2_generic(IndexType n, const X *x, IndexType incX, T &norm)
43 {
44 CXXBLAS_DEBUG_OUT("nrm2_generic");
45
46 using std::abs;
47 using std::pow;
48 using std::sqrt;
49
50 if (n<1) {
51 norm = T(0);
52 } else if (n==1) {
53 norm = abs(*x);
54 } else {
55 T scale = 0;
56 T ssq = 1;
57 // The following loop is equivalent to this call to the LAPACK
58 // auxiliary routine:
59 // CALL DLASSQ( N, X, INCX, SCALE, SSQ )
60 //
61 for (IndexType i=0, iX=0; i<n; ++i, iX+=incX) {
62 if (x[iX]!=T(0)) {
63 T absXi = abs(x[iX]);
64 if (scale<absXi) {
65 ssq = T(1) + ssq * pow(scale/absXi, 2);
66 scale = absXi;
67 } else {
68 ssq += pow(absXi/scale, 2);
69 }
70 }
71 }
72 norm = scale*sqrt(ssq);
73 }
74 }
75
76 template <typename IndexType, typename X, typename T>
77 void
78 nrm2(IndexType n, const X *x, IndexType incX, T &norm)
79 {
80 if (incX<0) {
81 x -= incX*(n-1);
82 }
83 nrm2_generic(n, x, incX, norm);
84 }
85
86 #ifdef HAVE_CBLAS
87
88 // snrm2
89 template <typename IndexType>
90 typename If<IndexType>::isBlasCompatibleInteger
91 nrm2(IndexType n, const float *x, IndexType incX, float &norm)
92 {
93 CXXBLAS_DEBUG_OUT("[" BLAS_IMPL "] cblas_snrm2");
94
95 norm = cblas_snrm2(n, x, incX);
96 }
97
98 // dnrm2
99 template <typename IndexType>
100 typename If<IndexType>::isBlasCompatibleInteger
101 nrm2(IndexType n, const double *x, IndexType incX, double &norm)
102 {
103 CXXBLAS_DEBUG_OUT("[" BLAS_IMPL "] cblas_dnrm2");
104
105 norm = cblas_dnrm2(n, x, incX);
106 }
107
108 // scnrm2
109 template <typename IndexType>
110 typename If<IndexType>::isBlasCompatibleInteger
111 nrm2(IndexType n, const ComplexFloat *x, IndexType incX, float &norm)
112 {
113 CXXBLAS_DEBUG_OUT("[" BLAS_IMPL "] cblas_scnrm2");
114
115 norm = cblas_scnrm2(n, x, incX);
116 }
117
118 // dznrm2
119 template <typename IndexType>
120 typename If<IndexType>::isBlasCompatibleInteger
121 nrm2(IndexType n, const ComplexDouble *x, IndexType incX, double &norm)
122 {
123 CXXBLAS_DEBUG_OUT("[" BLAS_IMPL "] cblas_dznrm2");
124
125 norm = cblas_dznrm2(n, x, incX);
126 }
127
128 #endif // HAVE_CBLAS
129
130 } // namespace cxxblas
131
132 #endif // CXXBLAS_LEVEL1_NRM2_TCC