3 * Copyright (C) 1996, 1997, 1998, 1999, 2000 Gerard Jungman
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 3 of the License, or (at
8 * your option) any later version.
10 * This program is distributed in the hope that it will be useful, but
11 * WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * General Public License for more details.
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
22 const int conj = (order == CblasColMajor) ? -1 : 1;
24 const BASE alpha_real = CONST_REAL0(alpha);
25 const BASE alpha_imag = CONST_IMAG0(alpha);
27 const BASE beta_real = CONST_REAL0(beta);
28 const BASE beta_imag = CONST_IMAG0(beta);
33 if ((alpha_real == 0.0 && alpha_imag == 0.0)
34 && (beta_real == 1.0 && beta_imag == 0.0))
37 /* form y := beta*y */
38 if (beta_real == 0.0 && beta_imag == 0.0) {
39 INDEX iy = OFFSET(N, incY);
40 for (i = 0; i < N; i++) {
45 } else if (!(beta_real == 1.0 && beta_imag == 0.0)) {
46 INDEX iy = OFFSET(N, incY);
47 for (i = 0; i < N; i++) {
48 const BASE y_real = REAL(Y, iy);
49 const BASE y_imag = IMAG(Y, iy);
50 const BASE tmpR = y_real * beta_real - y_imag * beta_imag;
51 const BASE tmpI = y_real * beta_imag + y_imag * beta_real;
58 if (alpha_real == 0.0 && alpha_imag == 0.0)
61 /* form y := alpha*A*x + y */
63 if ((order == CblasRowMajor && Uplo == CblasUpper)
64 || (order == CblasColMajor && Uplo == CblasLower)) {
65 INDEX ix = OFFSET(N, incX);
66 INDEX iy = OFFSET(N, incY);
67 for (i = 0; i < N; i++) {
68 BASE x_real = CONST_REAL(X, ix);
69 BASE x_imag = CONST_IMAG(X, ix);
70 BASE temp1_real = alpha_real * x_real - alpha_imag * x_imag;
71 BASE temp1_imag = alpha_real * x_imag + alpha_imag * x_real;
72 BASE temp2_real = 0.0;
73 BASE temp2_imag = 0.0;
74 const INDEX j_min = i + 1;
75 const INDEX j_max = GSL_MIN(N, i + K + 1);
76 INDEX jx = OFFSET(N, incX) + j_min * incX;
77 INDEX jy = OFFSET(N, incY) + j_min * incY;
78 BASE Aii_real = CONST_REAL(A, lda * i + 0);
79 /* Aii_imag is zero */
80 REAL(Y, iy) += temp1_real * Aii_real;
81 IMAG(Y, iy) += temp1_imag * Aii_real;
82 for (j = j_min; j < j_max; j++) {
83 BASE Aij_real = CONST_REAL(A, lda * i + (j - i));
84 BASE Aij_imag = conj * CONST_IMAG(A, lda * i + (j - i));
85 REAL(Y, jy) += temp1_real * Aij_real - temp1_imag * (-Aij_imag);
86 IMAG(Y, jy) += temp1_real * (-Aij_imag) + temp1_imag * Aij_real;
87 x_real = CONST_REAL(X, jx);
88 x_imag = CONST_IMAG(X, jx);
89 temp2_real += x_real * Aij_real - x_imag * Aij_imag;
90 temp2_imag += x_real * Aij_imag + x_imag * Aij_real;
94 REAL(Y, iy) += alpha_real * temp2_real - alpha_imag * temp2_imag;
95 IMAG(Y, iy) += alpha_real * temp2_imag + alpha_imag * temp2_real;
99 } else if ((order == CblasRowMajor && Uplo == CblasLower)
100 || (order == CblasColMajor && Uplo == CblasUpper)) {
101 INDEX ix = OFFSET(N, incX);
102 INDEX iy = OFFSET(N, incY);
103 for (i = 0; i < N; i++) {
104 BASE x_real = CONST_REAL(X, ix);
105 BASE x_imag = CONST_IMAG(X, ix);
106 BASE temp1_real = alpha_real * x_real - alpha_imag * x_imag;
107 BASE temp1_imag = alpha_real * x_imag + alpha_imag * x_real;
108 BASE temp2_real = 0.0;
109 BASE temp2_imag = 0.0;
110 const INDEX j_min = (K > i ? 0 : i - K);
111 const INDEX j_max = i;
112 INDEX jx = OFFSET(N, incX) + j_min * incX;
113 INDEX jy = OFFSET(N, incY) + j_min * incY;
115 for (j = j_min; j < j_max; j++) {
116 BASE Aij_real = CONST_REAL(A, i * lda + (K - i + j));
117 BASE Aij_imag = conj * CONST_IMAG(A, i * lda + (K - i + j));
118 REAL(Y, jy) += temp1_real * Aij_real - temp1_imag * (-Aij_imag);
119 IMAG(Y, jy) += temp1_real * (-Aij_imag) + temp1_imag * Aij_real;
120 x_real = CONST_REAL(X, jx);
121 x_imag = CONST_IMAG(X, jx);
122 temp2_real += x_real * Aij_real - x_imag * Aij_imag;
123 temp2_imag += x_real * Aij_imag + x_imag * Aij_real;
129 BASE Aii_real = CONST_REAL(A, lda * i + K);
130 /* Aii_imag is zero */
131 REAL(Y, iy) += temp1_real * Aii_real;
132 IMAG(Y, iy) += temp1_imag * Aii_real;
135 REAL(Y, iy) += alpha_real * temp2_real - alpha_imag * temp2_imag;
136 IMAG(Y, iy) += alpha_real * temp2_imag + alpha_imag * temp2_real;
142 BLAS_ERROR("unrecognized operation");