|
kusano |
2b45e8 |
/*********************************************************************/
|
|
kusano |
2b45e8 |
/* Copyright 2009, 2010 The University of Texas at Austin. */
|
|
kusano |
2b45e8 |
/* All rights reserved. */
|
|
kusano |
2b45e8 |
/* */
|
|
kusano |
2b45e8 |
/* Redistribution and use in source and binary forms, with or */
|
|
kusano |
2b45e8 |
/* without modification, are permitted provided that the following */
|
|
kusano |
2b45e8 |
/* conditions are met: */
|
|
kusano |
2b45e8 |
/* */
|
|
kusano |
2b45e8 |
/* 1. Redistributions of source code must retain the above */
|
|
kusano |
2b45e8 |
/* copyright notice, this list of conditions and the following */
|
|
kusano |
2b45e8 |
/* disclaimer. */
|
|
kusano |
2b45e8 |
/* */
|
|
kusano |
2b45e8 |
/* 2. Redistributions in binary form must reproduce the above */
|
|
kusano |
2b45e8 |
/* copyright notice, this list of conditions and the following */
|
|
kusano |
2b45e8 |
/* disclaimer in the documentation and/or other materials */
|
|
kusano |
2b45e8 |
/* provided with the distribution. */
|
|
kusano |
2b45e8 |
/* */
|
|
kusano |
2b45e8 |
/* THIS SOFTWARE IS PROVIDED BY THE UNIVERSITY OF TEXAS AT */
|
|
kusano |
2b45e8 |
/* AUSTIN ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, */
|
|
kusano |
2b45e8 |
/* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */
|
|
kusano |
2b45e8 |
/* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */
|
|
kusano |
2b45e8 |
/* DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY OF TEXAS AT */
|
|
kusano |
2b45e8 |
/* AUSTIN OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, */
|
|
kusano |
2b45e8 |
/* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES */
|
|
kusano |
2b45e8 |
/* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE */
|
|
kusano |
2b45e8 |
/* GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR */
|
|
kusano |
2b45e8 |
/* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF */
|
|
kusano |
2b45e8 |
/* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT */
|
|
kusano |
2b45e8 |
/* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT */
|
|
kusano |
2b45e8 |
/* OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE */
|
|
kusano |
2b45e8 |
/* POSSIBILITY OF SUCH DAMAGE. */
|
|
kusano |
2b45e8 |
/* */
|
|
kusano |
2b45e8 |
/* The views and conclusions contained in the software and */
|
|
kusano |
2b45e8 |
/* documentation are those of the authors and should not be */
|
|
kusano |
2b45e8 |
/* interpreted as representing official policies, either expressed */
|
|
kusano |
2b45e8 |
/* or implied, of The University of Texas at Austin. */
|
|
kusano |
2b45e8 |
/*********************************************************************/
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
#include <stdio.h></stdio.h>
|
|
kusano |
2b45e8 |
#include <ctype.h></ctype.h>
|
|
kusano |
2b45e8 |
#include "common.h"
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
const static FLOAT dp1 = 1.;
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
int CNAME(BLASLONG m, FLOAT *a, BLASLONG lda, FLOAT *b, BLASLONG incb, FLOAT *buffer){
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
BLASLONG i, is, min_i;
|
|
kusano |
2b45e8 |
FLOAT *gemvbuffer = (FLOAT *)buffer;
|
|
kusano |
2b45e8 |
FLOAT *B = b;
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
if (incb != 1) {
|
|
kusano |
2b45e8 |
B = buffer;
|
|
kusano |
2b45e8 |
gemvbuffer = (FLOAT *)(((BLASLONG)buffer + m * sizeof(FLOAT) + 4095) & ~4095);
|
|
kusano |
2b45e8 |
COPY_K(m, b, incb, buffer, 1);
|
|
kusano |
2b45e8 |
}
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
for (is = 0; is < m; is += DTB_ENTRIES){
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
min_i = MIN(m - is, DTB_ENTRIES);
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
#ifndef TRANSA
|
|
kusano |
2b45e8 |
if (is > 0){
|
|
kusano |
2b45e8 |
GEMV_N(is, min_i, 0, dp1,
|
|
kusano |
2b45e8 |
a + is * lda, lda,
|
|
kusano |
2b45e8 |
B + is, 1,
|
|
kusano |
2b45e8 |
B, 1, gemvbuffer);
|
|
kusano |
2b45e8 |
}
|
|
kusano |
2b45e8 |
#endif
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
for (i = 0; i < min_i; i++) {
|
|
kusano |
2b45e8 |
FLOAT *AA = a + is + (i + is) * lda;
|
|
kusano |
2b45e8 |
FLOAT *BB = B + is;
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
#ifndef TRANSA
|
|
kusano |
2b45e8 |
if (i > 0) AXPYU_K(i, 0, 0, BB[i], AA, 1, BB, 1, NULL, 0);
|
|
kusano |
2b45e8 |
#endif
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
#ifndef UNIT
|
|
kusano |
2b45e8 |
BB[i] *= AA[i];
|
|
kusano |
2b45e8 |
#endif
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
#ifdef TRANSA
|
|
kusano |
2b45e8 |
if (i < min_i - 1) BB[i] += DOTU_K(min_i - i - 1, AA + i + 1, 1, BB + i + 1, 1);
|
|
kusano |
2b45e8 |
#endif
|
|
kusano |
2b45e8 |
}
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
#ifdef TRANSA
|
|
kusano |
2b45e8 |
if (m - is > min_i){
|
|
kusano |
2b45e8 |
GEMV_T(m - is - min_i, min_i, 0, dp1,
|
|
kusano |
2b45e8 |
a + is + min_i + is * lda, lda,
|
|
kusano |
2b45e8 |
B + is + min_i, 1,
|
|
kusano |
2b45e8 |
B + is, 1, gemvbuffer);
|
|
kusano |
2b45e8 |
}
|
|
kusano |
2b45e8 |
#endif
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
}
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
if (incb != 1) {
|
|
kusano |
2b45e8 |
COPY_K(m, buffer, 1, b, incb);
|
|
kusano |
2b45e8 |
}
|
|
kusano |
2b45e8 |
|
|
kusano |
2b45e8 |
return 0;
|
|
kusano |
2b45e8 |
}
|
|
kusano |
2b45e8 |
|