Skip to content

Commit d45c53e

Browse files
committed
[OpenBLAS]: benchmark for her/her2 LEVEL2 functions
[description]: benchmark for her/her2 [solution]: added benchmark for her/her2, modified makefile in benchmark [dts]:
1 parent c284099 commit d45c53e

3 files changed

Lines changed: 474 additions & 0 deletions

File tree

benchmark/Makefile

Lines changed: 98 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -78,6 +78,8 @@ goto :: slinpack.goto dlinpack.goto clinpack.goto zlinpack.goto \
7878
chemm.goto zhemm.goto \
7979
cherk.goto zherk.goto \
8080
cher2k.goto zher2k.goto \
81+
cher.goto zher.goto \
82+
cher2.goto zher2.goto \
8183
sgemv.goto dgemv.goto cgemv.goto zgemv.goto \
8284
sspmv.goto dspmv.goto \
8385
strmv.goto dtrmv.goto ctrmv.goto ztrmv.goto \
@@ -117,6 +119,8 @@ acml :: slinpack.acml dlinpack.acml clinpack.acml zlinpack.acml \
117119
chemm.acml zhemm.acml \
118120
cherk.acml zherk.acml \
119121
cher2k.acml zher2k.acml \
122+
cher.acml zher.acml \
123+
cher2.acml zher2.acml \
120124
sgemv.acml dgemv.acml cgemv.acml zgemv.acml \
121125
strmv.acml dtrmv.acml ctrmv.acml ztrmv.acml \
122126
stpmv.acml dtpmv.acml ctpmv.acml ztpmv.acml \
@@ -156,6 +160,8 @@ atlas :: slinpack.atlas dlinpack.atlas clinpack.atlas zlinpack.atlas \
156160
chemm.atlas zhemm.atlas \
157161
cherk.atlas zherk.atlas \
158162
cher2k.atlas zher2k.atlas \
163+
cher.atlas zher.atlas \
164+
cher2.atlas zher2.atlas \
159165
sgemv.atlas dgemv.atlas cgemv.atlas zgemv.atlas \
160166
sspmv.atlas dspmv.atlas \
161167
strmv.atlas dtrmv.atlas ctrmv.atlas ztrmv.atlas \
@@ -195,6 +201,8 @@ mkl :: slinpack.mkl dlinpack.mkl clinpack.mkl zlinpack.mkl \
195201
chemm.mkl zhemm.mkl \
196202
cherk.mkl zherk.mkl \
197203
cher2k.mkl zher2k.mkl \
204+
cher.mkl zher.mkl \
205+
cher2.mkl zher2.mkl \
198206
sgemv.mkl dgemv.mkl cgemv.mkl zgemv.mkl \
199207
strmv.mkl dtrmv.mkl ctrmv.mkl ztrmv.mkl \
200208
stpmv.mkl dtpmv.mkl ctpmv.mkl ztpmv.mkl \
@@ -233,6 +241,8 @@ goto :: sgemm.goto dgemm.goto cgemm.goto zgemm.goto \
233241
chemm.goto zhemm.goto \
234242
cherk.goto zherk.goto \
235243
cher2k.goto zher2k.goto \
244+
cher.goto zher.goto \
245+
cher2.goto zher2.goto \
236246
sgemv.goto dgemv.goto cgemv.goto zgemv.goto \
237247
sspmv.goto dspmv.goto \
238248
strmv.goto dtrmv.goto ctrmv.goto ztrmv.goto \
@@ -278,6 +288,8 @@ acml :: slinpack.acml dlinpack.acml clinpack.acml zlinpack.acml \
278288
chemm.acml zhemm.acml \
279289
cherk.acml zherk.acml \
280290
cher2k.acml zher2k.acml \
291+
cher.acml zher.acml \
292+
cher2.acml zher2.acml \
281293
sgemv.acml dgemv.acml cgemv.acml zgemv.acml \
282294
strmv.acml dtrmv.acml ctrmv.acml ztrmv.acml \
283295
stpmv.acml dtpmv.acml ctpmv.acml ztpmv.acml \
@@ -317,6 +329,8 @@ atlas :: slinpack.atlas dlinpack.atlas clinpack.atlas zlinpack.atlas \
317329
chemm.atlas zhemm.atlas \
318330
cherk.atlas zherk.atlas \
319331
cher2k.atlas zher2k.atlas \
332+
cher.atlas zher.atlas \
333+
cher2.atlas zher2.atlas \
320334
sgemv.atlas dgemv.atlas cgemv.atlas zgemv.atlas \
321335
sspmv.atlas dspmv.atlas \
322336
strmv.atlas dtrmv.atlas ctrmv.atlas ztrmv.atlas \
@@ -358,6 +372,8 @@ mkl :: slinpack.mkl dlinpack.mkl clinpack.mkl zlinpack.mkl \
358372
chemm.mkl zhemm.mkl \
359373
cherk.mkl zherk.mkl \
360374
cher2k.mkl zher2k.mkl \
375+
cher.mkl zher.mkl \
376+
cher2.mkl zher2.mkl \
361377
sgemv.mkl dgemv.mkl cgemv.mkl zgemv.mkl \
362378
strmv.mkl dtrmv.mkl ctrmv.mkl ztrmv.mkl \
363379
stpmv.mkl dtpmv.mkl ctpmv.mkl ztpmv.mkl \
@@ -407,6 +423,8 @@ veclib :: slinpack.veclib dlinpack.veclib clinpack.veclib zlinpack.veclib \
407423
chemm.veclib zhemm.veclib \
408424
cherk.veclib zherk.veclib \
409425
cher2k.veclib zher2k.veclib \
426+
cher.veclib zher.veclib \
427+
cher2.veclib zher2.veclib \
410428
sgemv.veclib dgemv.veclib cgemv.veclib zgemv.veclib \
411429
strmv.veclib dtrmv.veclib ctrmv.veclib ztrmv.veclib \
412430
stpmv.veclib dtpmv.veclib ctpmv.veclib ztpmv.veclib \
@@ -1244,6 +1262,74 @@ zher2k.mkl : zher2k.$(SUFFIX)
12441262
zher2k.veclib : zher2k.$(SUFFIX)
12451263
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
12461264

1265+
##################################### Cher ####################################################
1266+
1267+
cher.goto : cher.$(SUFFIX) ../$(LIBNAME)
1268+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
1269+
1270+
cher.acml : cher.$(SUFFIX)
1271+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1272+
1273+
cher.atlas : cher.$(SUFFIX)
1274+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1275+
1276+
cher.mkl : cher.$(SUFFIX)
1277+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1278+
1279+
cher.veclib : cher.$(SUFFIX)
1280+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1281+
1282+
##################################### Zher ####################################################
1283+
1284+
zher.goto : zher.$(SUFFIX) ../$(LIBNAME)
1285+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
1286+
1287+
zher.acml : zher.$(SUFFIX)
1288+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1289+
1290+
zher.atlas : zher.$(SUFFIX)
1291+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1292+
1293+
zher.mkl : zher.$(SUFFIX)
1294+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1295+
1296+
zher.veclib : zher.$(SUFFIX)
1297+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1298+
1299+
##################################### Cher2 ####################################################
1300+
1301+
cher2.goto : cher2.$(SUFFIX) ../$(LIBNAME)
1302+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
1303+
1304+
cher2.acml : cher2.$(SUFFIX)
1305+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1306+
1307+
cher2.atlas : cher2.$(SUFFIX)
1308+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1309+
1310+
cher2.mkl : cher2.$(SUFFIX)
1311+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1312+
1313+
cher2.veclib : cher2.$(SUFFIX)
1314+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1315+
1316+
##################################### Zher2 ####################################################
1317+
1318+
zher2.goto : zher2.$(SUFFIX) ../$(LIBNAME)
1319+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
1320+
1321+
zher2.acml : zher2.$(SUFFIX)
1322+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1323+
1324+
zher2.atlas : zher2.$(SUFFIX)
1325+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1326+
1327+
zher2.mkl : zher2.$(SUFFIX)
1328+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1329+
1330+
zher2.veclib : zher2.$(SUFFIX)
1331+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1332+
12471333
##################################### Sgemv ####################################################
12481334
sgemv.goto : sgemv.$(SUFFIX) ../$(LIBNAME)
12491335
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
@@ -2868,6 +2954,18 @@ cher2k.$(SUFFIX) : her2k.c
28682954
zher2k.$(SUFFIX) : her2k.c
28692955
$(CC) $(CFLAGS) -c -DCOMPLEX -DDOUBLE -o $(@F) $^
28702956

2957+
cher.$(SUFFIX) : her.c
2958+
$(CC) $(CFLAGS) -c -DCOMPLEX -UDOUBLE -o $(@F) $^
2959+
2960+
zher.$(SUFFIX) : her.c
2961+
$(CC) $(CFLAGS) -c -DCOMPLEX -DDOUBLE -o $(@F) $^
2962+
2963+
cher2.$(SUFFIX) : her2.c
2964+
$(CC) $(CFLAGS) -c -DCOMPLEX -UDOUBLE -o $(@F) $^
2965+
2966+
zher2.$(SUFFIX) : her2.c
2967+
$(CC) $(CFLAGS) -c -DCOMPLEX -DDOUBLE -o $(@F) $^
2968+
28712969
sgemv.$(SUFFIX) : gemv.c
28722970
$(CC) $(CFLAGS) -c -UCOMPLEX -UDOUBLE -o $(@F) $^
28732971

benchmark/her.c

Lines changed: 186 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,186 @@
1+
/***************************************************************************
2+
Copyright (c) 2020, The OpenBLAS Project
3+
All rights reserved.
4+
Redistribution and use in source and binary forms, with or without
5+
modification, are permitted provided that the following conditions are
6+
met:
7+
1. Redistributions of source code must retain the above copyright
8+
notice, this list of conditions and the following disclaimer.
9+
2. Redistributions in binary form must reproduce the above copyright
10+
notice, this list of conditions and the following disclaimer in
11+
the documentation and/or other materials provided with the
12+
distribution.
13+
3. Neither the name of the OpenBLAS project nor the names of
14+
its contributors may be used to endorse or promote products
15+
derived from this software without specific prior written permission.
16+
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
17+
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18+
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19+
ARE DISCLAIMED. IN NO EVENT SHALL THE OPENBLAS PROJECT OR CONTRIBUTORS BE
20+
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21+
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
22+
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
23+
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
24+
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
25+
USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26+
*****************************************************************************/
27+
28+
#include <stdio.h>
29+
#include <stdlib.h>
30+
#ifdef __CYGWIN32__
31+
#include <sys/time.h>
32+
#endif
33+
#include "common.h"
34+
35+
36+
#undef HER
37+
38+
39+
#ifdef DOUBLE
40+
#define HER BLASFUNC(zher)
41+
#else
42+
#define HER BLASFUNC(cher)
43+
#endif
44+
45+
46+
#if defined(__WIN32__) || defined(__WIN64__)
47+
48+
#ifndef DELTA_EPOCH_IN_MICROSECS
49+
#define DELTA_EPOCH_IN_MICROSECS 11644473600000000ULL
50+
#endif
51+
52+
int gettimeofday(struct timeval *tv, void *tz){
53+
54+
FILETIME ft;
55+
unsigned __int64 tmpres = 0;
56+
static int tzflag;
57+
58+
if (NULL != tv)
59+
{
60+
GetSystemTimeAsFileTime(&ft);
61+
62+
tmpres |= ft.dwHighDateTime;
63+
tmpres <<= 32;
64+
tmpres |= ft.dwLowDateTime;
65+
66+
/*converting file time to unix epoch*/
67+
tmpres /= 10; /*convert into microseconds*/
68+
tmpres -= DELTA_EPOCH_IN_MICROSECS;
69+
tv->tv_sec = (long)(tmpres / 1000000UL);
70+
tv->tv_usec = (long)(tmpres % 1000000UL);
71+
}
72+
73+
return 0;
74+
}
75+
76+
#endif
77+
78+
#if !defined(__WIN32__) && !defined(__WIN64__) && !defined(__CYGWIN32__) && 0
79+
80+
static void *huge_malloc(BLASLONG size){
81+
int shmid;
82+
void *address;
83+
84+
#ifndef SHM_HUGETLB
85+
#define SHM_HUGETLB 04000
86+
#endif
87+
88+
if ((shmid =shmget(IPC_PRIVATE,
89+
(size + HUGE_PAGESIZE) & ~(HUGE_PAGESIZE - 1),
90+
SHM_HUGETLB | IPC_CREAT |0600)) < 0) {
91+
printf( "Memory allocation failed(shmget).\n");
92+
exit(1);
93+
}
94+
95+
address = shmat(shmid, NULL, SHM_RND);
96+
97+
if ((BLASLONG)address == -1){
98+
printf( "Memory allocation failed(shmat).\n");
99+
exit(1);
100+
}
101+
102+
shmctl(shmid, IPC_RMID, 0);
103+
104+
return address;
105+
}
106+
107+
#define malloc huge_malloc
108+
109+
#endif
110+
111+
int main(int argc, char *argv[]){
112+
113+
FLOAT *a, *x;
114+
FLOAT alpha[] = {1.0, 1.0};
115+
blasint incx = 1;
116+
char *p;
117+
118+
char uplo='U';
119+
char trans='N';
120+
121+
if ((p = getenv("OPENBLAS_UPLO"))) uplo=*p;
122+
if ((p = getenv("OPENBLAS_TRANS"))) trans=*p;
123+
124+
blasint m, i, j;
125+
126+
int from = 1;
127+
int to = 200;
128+
int step = 1;
129+
130+
struct timeval start, stop;
131+
double time1;
132+
133+
argc--;argv++;
134+
135+
if (argc > 0) { from = atol(*argv); argc--; argv++;}
136+
if (argc > 0) { to = MAX(atol(*argv), from); argc--; argv++;}
137+
if (argc > 0) { step = atol(*argv); argc--; argv++;}
138+
139+
fprintf(stderr, "From : %3d To : %3d Step = %3d Uplo = %c Trans = %c\n", from, to, step,uplo,trans);
140+
141+
142+
if (( a = (FLOAT *)malloc(sizeof(FLOAT) * to * to * COMPSIZE)) == NULL){
143+
fprintf(stderr,"Out of Memory!!\n");exit(1);
144+
}
145+
146+
if (( x = (FLOAT *)malloc(sizeof(FLOAT) * to * COMPSIZE)) == NULL){
147+
fprintf(stderr,"Out of Memory!!\n");exit(1);
148+
}
149+
150+
151+
152+
#ifdef linux
153+
srandom(getpid());
154+
#endif
155+
156+
fprintf(stderr, " SIZE Flops\n");
157+
158+
for(m = from; m <= to; m += step)
159+
{
160+
fprintf(stderr, " %6d : ", (int)m);
161+
162+
for(j = 0; j < m; j++){
163+
for(i = 0; i < m * COMPSIZE; i++){
164+
a[(long)i + (long)j * (long)m * COMPSIZE] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
165+
}
166+
x[ (long)j * COMPSIZE] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
167+
}
168+
169+
gettimeofday( &start, (struct timezone *)0);
170+
171+
HER (&uplo, &m, alpha, x, &incx, a, &m );
172+
173+
gettimeofday( &stop, (struct timezone *)0);
174+
175+
time1 = (double)(stop.tv_sec - start.tv_sec) + (double)((stop.tv_usec - start.tv_usec)) * 1.e-6;
176+
177+
gettimeofday( &start, (struct timezone *)0);
178+
179+
fprintf(stderr,
180+
" %10.2f MFlops\n",
181+
COMPSIZE * COMPSIZE * 1. * (double)m * (double)m / time1 * 1.e-6);
182+
183+
}
184+
185+
return 0;
186+
}

0 commit comments

Comments
 (0)