Skip to content

Commit 18099de

Browse files
authored
Merge pull request #2487 from jijiwawa/develop
add benchmark for spr/spr2
2 parents 9f5a74f + 97c36ca commit 18099de

3 files changed

Lines changed: 499 additions & 0 deletions

File tree

benchmark/Makefile

Lines changed: 94 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -56,6 +56,8 @@ goto :: slinpack.goto dlinpack.goto clinpack.goto zlinpack.goto \
5656
sgemm.goto dgemm.goto cgemm.goto zgemm.goto \
5757
strmm.goto dtrmm.goto ctrmm.goto ztrmm.goto \
5858
strsm.goto dtrsm.goto ctrsm.goto ztrsm.goto \
59+
sspr.goto dspr.goto \
60+
sspr2.goto dspr2.goto \
5961
ssyr.goto dsyr.goto \
6062
ssyr2.goto dsyr2.goto \
6163
ssyrk.goto dsyrk.goto csyrk.goto zsyrk.goto \
@@ -87,6 +89,8 @@ acml :: slinpack.acml dlinpack.acml clinpack.acml zlinpack.acml \
8789
sgemm.acml dgemm.acml cgemm.acml zgemm.acml \
8890
strmm.acml dtrmm.acml ctrmm.acml ztrmm.acml \
8991
strsm.acml dtrsm.acml ctrsm.acml ztrsm.acml \
92+
sspr.acml dspr.acml \
93+
sspr2.acml dspr2.acml \
9094
ssyr.acml dsyr.acml \
9195
ssyr2.acml dsyr2.acml \
9296
ssyrk.acml dsyrk.acml csyrk.acml zsyrk.acml \
@@ -117,6 +121,8 @@ atlas :: slinpack.atlas dlinpack.atlas clinpack.atlas zlinpack.atlas \
117121
sgemm.atlas dgemm.atlas cgemm.atlas zgemm.atlas \
118122
strmm.atlas dtrmm.atlas ctrmm.atlas ztrmm.atlas \
119123
strsm.atlas dtrsm.atlas ctrsm.atlas ztrsm.atlas \
124+
sspr.atlas dspr.atlas \
125+
sspr2.atlas dspr2.atlas \
120126
ssyr.atlas dsyr.atlas \
121127
ssyr2.atlas dsyr2.atlas \
122128
ssyrk.atlas dsyrk.atlas csyrk.atlas zsyrk.atlas \
@@ -148,6 +154,8 @@ mkl :: slinpack.mkl dlinpack.mkl clinpack.mkl zlinpack.mkl \
148154
sgemm.mkl dgemm.mkl cgemm.mkl zgemm.mkl \
149155
strmm.mkl dtrmm.mkl ctrmm.mkl ztrmm.mkl \
150156
strsm.mkl dtrsm.mkl ctrsm.mkl ztrsm.mkl \
157+
sspr.mkl dspr.mkl \
158+
sspr2.mkl dspr2.mkl \
151159
ssyr.mkl dsyr.mkl \
152160
ssyr2.mkl dsyr2.mkl \
153161
ssyrk.mkl dsyrk.mkl csyrk.mkl zsyrk.mkl \
@@ -178,6 +186,8 @@ else
178186
goto :: sgemm.goto dgemm.goto cgemm.goto zgemm.goto \
179187
strmm.goto dtrmm.goto ctrmm.goto ztrmm.goto \
180188
strsm.goto dtrsm.goto ctrsm.goto ztrsm.goto \
189+
sspr.goto dspr.goto \
190+
sspr2.goto dspr2.goto \
181191
ssyr.goto dsyr.goto \
182192
ssyr2.goto dsyr2.goto \
183193
ssyrk.goto dsyrk.goto csyrk.goto zsyrk.goto \
@@ -208,6 +218,8 @@ acml :: slinpack.acml dlinpack.acml clinpack.acml zlinpack.acml \
208218
sgemm.acml dgemm.acml cgemm.acml zgemm.acml \
209219
strmm.acml dtrmm.acml ctrmm.acml ztrmm.acml \
210220
strsm.acml dtrsm.acml ctrsm.acml ztrsm.acml \
221+
sspr.acml dspr.acml \
222+
sspr2.acml dspr2.acml \
211223
ssyr.acml dsyr.acml \
212224
ssyr2.acml dsyr2.acml \
213225
ssyrk.acml dsyrk.acml csyrk.acml zsyrk.acml \
@@ -238,6 +250,8 @@ atlas :: slinpack.atlas dlinpack.atlas clinpack.atlas zlinpack.atlas \
238250
sgemm.atlas dgemm.atlas cgemm.atlas zgemm.atlas \
239251
strmm.atlas dtrmm.atlas ctrmm.atlas ztrmm.atlas \
240252
strsm.atlas dtrsm.atlas ctrsm.atlas ztrsm.atlas \
253+
sspr.atlas dspr.atlas \
254+
sspr2.atlas dspr2.atlas \
241255
ssyr.atlas dsyr.atlas \
242256
ssyr2.atlas dsyr2.atlas \
243257
ssyrk.atlas dsyrk.atlas csyrk.atlas zsyrk.atlas \
@@ -271,6 +285,8 @@ mkl :: slinpack.mkl dlinpack.mkl clinpack.mkl zlinpack.mkl \
271285
sgemm.mkl dgemm.mkl cgemm.mkl zgemm.mkl \
272286
strmm.mkl dtrmm.mkl ctrmm.mkl ztrmm.mkl \
273287
strsm.mkl dtrsm.mkl ctrsm.mkl ztrsm.mkl \
288+
sspr.mkl dspr.mkl \
289+
sspr2.mkl dspr2.mkl \
274290
ssyr.mkl dsyr.mkl \
275291
ssyr2.mkl dsyr2.mkl \
276292
ssyrk.mkl dsyrk.mkl csyrk.mkl zsyrk.mkl \
@@ -312,6 +328,8 @@ veclib :: slinpack.veclib dlinpack.veclib clinpack.veclib zlinpack.veclib \
312328
sgemm.veclib dgemm.veclib cgemm.veclib zgemm.veclib \
313329
strmm.veclib dtrmm.veclib ctrmm.veclib ztrmm.veclib \
314330
strsm.veclib dtrsm.veclib ctrsm.veclib ztrsm.veclib \
331+
sspr.veclib dspr.veclib \
332+
sspr2.veclib dspr2.veclib \
315333
ssyr.veclib dsyr.veclib \
316334
ssyr2.veclib dsyr2.veclib \
317335
ssyrk.veclib dsyrk.veclib csyrk.veclib zsyrk.veclib \
@@ -834,6 +852,70 @@ dsyr.mkl : dsyr.$(SUFFIX)
834852

835853
dsyr.veclib : dsyr.$(SUFFIX)
836854
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
855+
856+
##################################### Sspr ####################################################
857+
sspr.goto : sspr.$(SUFFIX) ../$(LIBNAME)
858+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
859+
860+
sspr.acml : sspr.$(SUFFIX)
861+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
862+
863+
sspr.atlas : sspr.$(SUFFIX)
864+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
865+
866+
sspr.mkl : sspr.$(SUFFIX)
867+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
868+
869+
sspr.veclib : sspr.$(SUFFIX)
870+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
871+
872+
##################################### Dspr ####################################################
873+
dspr.goto : dspr.$(SUFFIX) ../$(LIBNAME)
874+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
875+
876+
dspr.acml : dspr.$(SUFFIX)
877+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
878+
879+
dspr.atlas : dspr.$(SUFFIX)
880+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
881+
882+
dspr.mkl : dspr.$(SUFFIX)
883+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
884+
885+
dspr.veclib : dspr.$(SUFFIX)
886+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
887+
888+
##################################### Sspr2 ####################################################
889+
sspr2.goto : sspr2.$(SUFFIX) ../$(LIBNAME)
890+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
891+
892+
sspr2.acml : sspr2.$(SUFFIX)
893+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
894+
895+
sspr2.atlas : sspr2.$(SUFFIX)
896+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
897+
898+
sspr2.mkl : sspr2.$(SUFFIX)
899+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
900+
901+
sspr2.veclib : sspr2.$(SUFFIX)
902+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
903+
904+
##################################### Dspr2 ####################################################
905+
dspr2.goto : dspr2.$(SUFFIX) ../$(LIBNAME)
906+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
907+
908+
dspr2.acml : dspr2.$(SUFFIX)
909+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
910+
911+
dspr2.atlas : dspr2.$(SUFFIX)
912+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
913+
914+
dspr2.mkl : dspr2.$(SUFFIX)
915+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
916+
917+
dspr2.veclib : dspr2.$(SUFFIX)
918+
837919
##################################### Ssyr2 ####################################################
838920
ssyr2.goto : ssyr2.$(SUFFIX) ../$(LIBNAME)
839921
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
@@ -2311,6 +2393,18 @@ ssyr.$(SUFFIX) : syr.c
23112393

23122394
dsyr.$(SUFFIX) : syr.c
23132395
$(CC) $(CFLAGS) -c -UCOMPLEX -DDOUBLE -o $(@F) $^
2396+
2397+
sspr.$(SUFFIX) : spr.c
2398+
$(CC) $(CFLAGS) -c -UCOMPLEX -UDOUBLE -o $(@F) $^
2399+
2400+
dspr.$(SUFFIX) : spr.c
2401+
$(CC) $(CFLAGS) -c -UCOMPLEX -DDOUBLE -o $(@F) $^
2402+
2403+
sspr2.$(SUFFIX) : spr2.c
2404+
$(CC) $(CFLAGS) -c -UCOMPLEX -UDOUBLE -o $(@F) $^
2405+
2406+
dspr2.$(SUFFIX) : spr2.c
2407+
$(CC) $(CFLAGS) -c -UCOMPLEX -DDOUBLE -o $(@F) $^
23142408

23152409
ssyr2.$(SUFFIX) : syr2.c
23162410
$(CC) $(CFLAGS) -c -UCOMPLEX -UDOUBLE -o $(@F) $^

benchmark/spr.c

Lines changed: 198 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,198 @@
1+
/***************************************************************************
2+
Copyright (c) 2014, 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 SPR
37+
38+
#ifdef DOUBLE
39+
#define SPR BLASFUNC(dspr)
40+
#else
41+
#define SPR BLASFUNC(sspr)
42+
#endif
43+
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,*c;
114+
FLOAT alpha[] = {1.0, 1.0};
115+
blasint inc_x=1;
116+
int loops = 1;
117+
int l;
118+
char *p;
119+
120+
char uplo='U';
121+
122+
if ((p = getenv("OPENBLAS_UPLO"))) uplo=*p;
123+
if ((p = getenv("OPENBLAS_LOOPS"))) loops = atoi(p);
124+
if ((p = getenv("OPENBLAS_INCX"))) inc_x = atoi(p);
125+
126+
blasint m, i, j;
127+
128+
int from = 1;
129+
int to = 200;
130+
int step = 1;
131+
132+
struct timeval start, stop;
133+
double time1,timeg;
134+
135+
argc--;argv++;
136+
137+
if (argc > 0) { from = atol(*argv); argc--; argv++;}
138+
if (argc > 0) { to = MAX(atol(*argv), from); argc--; argv++;}
139+
if (argc > 0) { step = atol(*argv); argc--; argv++;}
140+
141+
fprintf(stderr, "From : %3d To : %3d Step = %3d Uplo = %c Inc_x = %d\n", from, to, step,uplo,inc_x);
142+
143+
144+
if (( a = (FLOAT *)malloc(sizeof(FLOAT) * to * to * COMPSIZE)) == NULL){
145+
fprintf(stderr,"Out of Memory!!\n");exit(1);
146+
}
147+
148+
if (( c = (FLOAT *)malloc(sizeof(FLOAT) * to * abs(inc_x) * COMPSIZE)) == NULL){
149+
fprintf(stderr,"Out of Memory!!\n");exit(1);
150+
}
151+
152+
#ifdef linux
153+
srandom(getpid());
154+
#endif
155+
156+
fprintf(stderr, " SIZE Flops Time\n");
157+
158+
for(m = from; m <= to; m += step)
159+
{
160+
timeg=0;
161+
162+
fprintf(stderr, " %6d : ", (int)m);
163+
164+
for (l=0; l<loops; l++)
165+
{
166+
167+
for(j = 0; j < m; j++){
168+
for(i = 0; i < m * COMPSIZE; i++){
169+
a[(long)i + (long)j * (long)m * COMPSIZE] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
170+
}
171+
}
172+
for(i = 0; i < m * COMPSIZE * abs(inc_x); i++){
173+
c[i] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
174+
}
175+
176+
gettimeofday( &start, (struct timezone *)0);
177+
178+
SPR (&uplo, &m, alpha, c, &inc_x, a);
179+
180+
gettimeofday( &stop, (struct timezone *)0);
181+
182+
time1 = (double)(stop.tv_sec - start.tv_sec) + (double)((stop.tv_usec - start.tv_usec)) * 1.e-6;
183+
184+
timeg += time1;
185+
}
186+
187+
timeg /= loops;
188+
189+
fprintf(stderr,
190+
" %10.2f MBytes %10.6f sec\n",
191+
COMPSIZE * COMPSIZE * 1. * (double)m * (double)m / timeg * 1.e-6, timeg);
192+
193+
}
194+
195+
return 0;
196+
}
197+
198+
// void main(int argc, char *argv[]) __attribute__((weak, alias("MAIN__")));

0 commit comments

Comments
 (0)