1 /*********************************************************************/
2 /* Copyright 2009, 2010 The University of Texas at Austin. */
3 /* All rights reserved. */
5 /* Redistribution and use in source and binary forms, with or */
6 /* without modification, are permitted provided that the following */
7 /* conditions are met: */
9 /* 1. Redistributions of source code must retain the above */
10 /* copyright notice, this list of conditions and the following */
13 /* 2. Redistributions in binary form must reproduce the above */
14 /* copyright notice, this list of conditions and the following */
15 /* disclaimer in the documentation and/or other materials */
16 /* provided with the distribution. */
18 /* THIS SOFTWARE IS PROVIDED BY THE UNIVERSITY OF TEXAS AT */
19 /* AUSTIN ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, */
20 /* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */
21 /* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */
22 /* DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY OF TEXAS AT */
23 /* AUSTIN OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, */
24 /* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES */
25 /* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE */
26 /* GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR */
27 /* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF */
28 /* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT */
29 /* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT */
30 /* OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE */
31 /* POSSIBILITY OF SUCH DAMAGE. */
33 /* The views and conclusions contained in the software and */
34 /* documentation are those of the authors and should not be */
35 /* interpreted as representing official policies, either expressed */
36 /* or implied, of The University of Texas at Austin. */
37 /*********************************************************************/
42 #ifdef FUNCTION_PROFILE
43 #include "functable.h"
47 #define ERROR_NAME "XHPR2 "
49 #define ERROR_NAME "ZHPR2 "
51 #define ERROR_NAME "CHPR2 "
54 static int (*hpr2[])(BLASLONG, FLOAT, FLOAT, FLOAT *, BLASLONG, FLOAT *, BLASLONG, FLOAT *, FLOAT *) = {
56 xhpr2_U, xhpr2_L, xhpr2_V, xhpr2_M,
58 zhpr2_U, zhpr2_L, zhpr2_V, zhpr2_M,
60 chpr2_U, chpr2_L, chpr2_V, chpr2_M,
65 static int (*hpr2_thread[])(BLASLONG, FLOAT *, FLOAT *, BLASLONG, FLOAT *, BLASLONG, FLOAT *, FLOAT *, int) = {
67 xhpr2_thread_U, xhpr2_thread_L, xhpr2_thread_V, xhpr2_thread_M,
69 zhpr2_thread_U, zhpr2_thread_L, zhpr2_thread_V, zhpr2_thread_M,
71 chpr2_thread_U, chpr2_thread_L, chpr2_thread_V, chpr2_thread_M,
78 void NAME(char *UPLO, blasint *N, FLOAT *ALPHA,
79 FLOAT *x, blasint *INCX, FLOAT *y, blasint *INCY, FLOAT *a){
81 char uplo_arg = *UPLO;
83 FLOAT alpha_r = ALPHA[0];
84 FLOAT alpha_i = ALPHA[1];
100 if (uplo_arg == 'U') uplo = 0;
101 if (uplo_arg == 'L') uplo = 1;
105 if (incy == 0) info = 7;
106 if (incx == 0) info = 5;
108 if (uplo < 0) info = 1;
111 BLASFUNC(xerbla)(ERROR_NAME, &info, sizeof(ERROR_NAME));
117 void CNAME(enum CBLAS_ORDER order,
118 enum CBLAS_UPLO Uplo,
121 void *vx, blasint incx,
122 void *vy, blasint incy,
125 FLOAT* ALPHA = (FLOAT*) VALPHA;
126 FLOAT* x = (FLOAT*) vx;
127 FLOAT* y = (FLOAT*) vy;
128 FLOAT* a = (FLOAT*) va;
130 FLOAT alpha_r = ALPHA[0];
131 FLOAT alpha_i = ALPHA[1];
144 if (order == CblasColMajor) {
145 if (Uplo == CblasUpper) uplo = 0;
146 if (Uplo == CblasLower) uplo = 1;
150 if (incy == 0) info = 7;
151 if (incx == 0) info = 5;
153 if (uplo < 0) info = 1;
156 if (order == CblasRowMajor) {
157 if (Uplo == CblasUpper) uplo = 3;
158 if (Uplo == CblasLower) uplo = 2;
162 if (incx == 0) info = 7;
163 if (incy == 0) info = 5;
165 if (uplo < 0) info = 1;
169 BLASFUNC(xerbla)(ERROR_NAME, &info, sizeof(ERROR_NAME));
177 if ((alpha_r == ZERO) && (alpha_i == ZERO)) return;
182 FUNCTION_PROFILE_START();
184 if (incx < 0 ) x -= (n - 1) * incx * 2;
185 if (incy < 0 ) y -= (n - 1) * incy * 2;
187 buffer = (FLOAT *)blas_memory_alloc(1);
190 nthreads = num_cpu_avail(2);
195 (hpr2[uplo])(n, alpha_r, alpha_i, x, incx, y, incy, a, buffer);
200 (hpr2_thread[uplo])(n, ALPHA, x, incx, y, incy, a, buffer, nthreads);
205 blas_memory_free(buffer);
207 FUNCTION_PROFILE_END(4, n * n / 2 + 2 * n, 2 * n * n);