Skip to content

Commit 0f1a2b1

Browse files
committed
add benchmark for spr/spr2
1 parent ad9e531 commit 0f1a2b1

3 files changed

Lines changed: 500 additions & 1 deletion

File tree

benchmark/Makefile

Lines changed: 95 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -57,6 +57,8 @@ goto :: slinpack.goto dlinpack.goto clinpack.goto zlinpack.goto \
5757
strmm.goto dtrmm.goto ctrmm.goto ztrmm.goto \
5858
strsm.goto dtrsm.goto ctrsm.goto ztrsm.goto \
5959
ssyr.goto dsyr.goto \
60+
sspr.goto dspr.goto \
61+
sspr2.goto dspr2.goto \
6062
ssyrk.goto dsyrk.goto csyrk.goto zsyrk.goto \
6163
ssyr2k.goto dsyr2k.goto csyr2k.goto zsyr2k.goto \
6264
sger.goto dger.goto cger.goto zger.goto \
@@ -86,6 +88,8 @@ acml :: slinpack.acml dlinpack.acml clinpack.acml zlinpack.acml \
8688
strmm.acml dtrmm.acml ctrmm.acml ztrmm.acml \
8789
strsm.acml dtrsm.acml ctrsm.acml ztrsm.acml \
8890
ssyr.acml dsyr.acml \
91+
sspr.acml dspr.acml \
92+
sspr2.acml dspr2.acml \
8993
ssyrk.acml dsyrk.acml csyrk.acml zsyrk.acml \
9094
ssyr2k.acml dsyr2k.acml csyr2k.acml zsyr2k.acml \
9195
sger.acml dger.acml cger.acml zger.acml \
@@ -113,7 +117,9 @@ atlas :: slinpack.atlas dlinpack.atlas clinpack.atlas zlinpack.atlas \
113117
sgemm.atlas dgemm.atlas cgemm.atlas zgemm.atlas \
114118
strmm.atlas dtrmm.atlas ctrmm.atlas ztrmm.atlas \
115119
strsm.atlas dtrsm.atlas ctrsm.atlas ztrsm.atlas \
116-
ssyr.goto dsyr.atlas \
120+
ssyr.atlas dsyr.atlas \
121+
sspr.atlas dspr.atlas \
122+
sspr2.atlas dspr2.atlas \
117123
ssyrk.atlas dsyrk.atlas csyrk.atlas zsyrk.atlas \
118124
ssyr2k.atlas dsyr2k.atlas csyr2k.atlas zsyr2k.atlas \
119125
sger.atlas dger.atlas cger.atlas zger.atlas\
@@ -143,6 +149,8 @@ mkl :: slinpack.mkl dlinpack.mkl clinpack.mkl zlinpack.mkl \
143149
strmm.mkl dtrmm.mkl ctrmm.mkl ztrmm.mkl \
144150
strsm.mkl dtrsm.mkl ctrsm.mkl ztrsm.mkl \
145151
ssyr.mkl dsyr.mkl \
152+
sspr.mkl dspr.mkl \
153+
sspr2.mkl dspr2.mkl \
146154
ssyrk.mkl dsyrk.mkl csyrk.mkl zsyrk.mkl \
147155
ssyr2k.mkl dsyr2k.mkl csyr2k.mkl zsyr2k.mkl \
148156
sger.mkl dger.mkl cger.mkl zger.mkl \
@@ -171,6 +179,8 @@ goto :: sgemm.goto dgemm.goto cgemm.goto zgemm.goto \
171179
strmm.goto dtrmm.goto ctrmm.goto ztrmm.goto \
172180
strsm.goto dtrsm.goto ctrsm.goto ztrsm.goto \
173181
ssyr.goto dsyr.goto \
182+
sspr.goto dspr.goto \
183+
sspr2.goto dspr2.goto \
174184
ssyrk.goto dsyrk.goto csyrk.goto zsyrk.goto \
175185
ssyr2k.goto dsyr2k.goto csyr2k.goto zsyr2k.goto \
176186
sger.goto dger.goto cger.goto zger.goto \
@@ -199,6 +209,8 @@ acml :: slinpack.acml dlinpack.acml clinpack.acml zlinpack.acml \
199209
strmm.acml dtrmm.acml ctrmm.acml ztrmm.acml \
200210
strsm.acml dtrsm.acml ctrsm.acml ztrsm.acml \
201211
ssyr.acml dsyr.acml \
212+
sspr.acml dspr.acml \
213+
sspr2.acml dspr2.acml \
202214
ssyrk.acml dsyrk.acml csyrk.acml zsyrk.acml \
203215
ssyr2k.acml dsyr2k.acml csyr2k.acml zsyr2k.acml \
204216
sger.acml dger.acml cger.acml zger.acml \
@@ -227,6 +239,8 @@ atlas :: slinpack.atlas dlinpack.atlas clinpack.atlas zlinpack.atlas \
227239
strmm.atlas dtrmm.atlas ctrmm.atlas ztrmm.atlas \
228240
strsm.atlas dtrsm.atlas ctrsm.atlas ztrsm.atlas \
229241
ssyr.atlas dsyr.atlas \
242+
sspr.atlas dspr.atlas \
243+
sspr2.atlas dspr2.atlas \
230244
ssyrk.atlas dsyrk.atlas csyrk.atlas zsyrk.atlas \
231245
ssyr2k.atlas dsyr2k.atlas csyr2k.atlas zsyr2k.atlas \
232246
sger.atlas dger.atlas cger.atlas zger.atlas\
@@ -258,6 +272,8 @@ mkl :: slinpack.mkl dlinpack.mkl clinpack.mkl zlinpack.mkl \
258272
strmm.mkl dtrmm.mkl ctrmm.mkl ztrmm.mkl \
259273
strsm.mkl dtrsm.mkl ctrsm.mkl ztrsm.mkl \
260274
ssyr.mkl dsyr.mkl \
275+
sspr.mkl dspr.mkl \
276+
sspr2.mkl dspr2.mkl \
261277
ssyrk.mkl dsyrk.mkl csyrk.mkl zsyrk.mkl \
262278
ssyr2k.mkl dsyr2k.mkl csyr2k.mkl zsyr2k.mkl \
263279
sger.mkl dger.mkl cger.mkl zger.mkl \
@@ -297,6 +313,8 @@ veclib :: slinpack.veclib dlinpack.veclib clinpack.veclib zlinpack.veclib \
297313
strmm.veclib dtrmm.veclib ctrmm.veclib ztrmm.veclib \
298314
strsm.veclib dtrsm.veclib ctrsm.veclib ztrsm.veclib \
299315
ssyr.veclib dsyr.veclib \
316+
sspr.veclib dspr.veclib \
317+
sspr2.veclib dspr2.veclib \
300318
ssyrk.veclib dsyrk.veclib csyrk.veclib zsyrk.veclib \
301319
ssyr2k.veclib dsyr2k.veclib csyr2k.veclib zsyr2k.veclib \
302320
sger.veclib dger.veclib cger.veclib zger.veclib \
@@ -816,6 +834,70 @@ dsyr.mkl : dsyr.$(SUFFIX)
816834

817835
dsyr.veclib : dsyr.$(SUFFIX)
818836
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
837+
838+
##################################### Sspr ####################################################
839+
sspr.goto : sspr.$(SUFFIX) ../$(LIBNAME)
840+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
841+
842+
sspr.acml : sspr.$(SUFFIX)
843+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
844+
845+
sspr.atlas : sspr.$(SUFFIX)
846+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
847+
848+
sspr.mkl : sspr.$(SUFFIX)
849+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
850+
851+
sspr.veclib : sspr.$(SUFFIX)
852+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
853+
854+
##################################### Dspr ####################################################
855+
dspr.goto : dspr.$(SUFFIX) ../$(LIBNAME)
856+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
857+
858+
dspr.acml : dspr.$(SUFFIX)
859+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
860+
861+
dspr.atlas : dspr.$(SUFFIX)
862+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
863+
864+
dspr.mkl : dspr.$(SUFFIX)
865+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
866+
867+
dspr.veclib : dspr.$(SUFFIX)
868+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
869+
870+
##################################### Sspr2 ####################################################
871+
sspr2.goto : sspr2.$(SUFFIX) ../$(LIBNAME)
872+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
873+
874+
sspr2.acml : sspr2.$(SUFFIX)
875+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
876+
877+
sspr2.atlas : sspr2.$(SUFFIX)
878+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
879+
880+
sspr2.mkl : sspr2.$(SUFFIX)
881+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
882+
883+
sspr2.veclib : sspr2.$(SUFFIX)
884+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
885+
886+
##################################### Dspr2 ####################################################
887+
dspr2.goto : dspr2.$(SUFFIX) ../$(LIBNAME)
888+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
889+
890+
dspr2.acml : dspr2.$(SUFFIX)
891+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
892+
893+
dspr2.atlas : dspr2.$(SUFFIX)
894+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
895+
896+
dspr2.mkl : dspr2.$(SUFFIX)
897+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
898+
899+
dspr2.veclib : dspr2.$(SUFFIX)
900+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
819901

820902
##################################### Ssyrk ####################################################
821903
ssyrk.goto : ssyrk.$(SUFFIX) ../$(LIBNAME)
@@ -2197,6 +2279,18 @@ ssyr.$(SUFFIX) : syr.c
21972279

21982280
dsyr.$(SUFFIX) : syr.c
21992281
$(CC) $(CFLAGS) -c -UCOMPLEX -DDOUBLE -o $(@F) $^
2282+
2283+
sspr.$(SUFFIX) : spr.c
2284+
$(CC) $(CFLAGS) -c -UCOMPLEX -UDOUBLE -o $(@F) $^
2285+
2286+
dspr.$(SUFFIX) : spr.c
2287+
$(CC) $(CFLAGS) -c -UCOMPLEX -DDOUBLE -o $(@F) $^
2288+
2289+
sspr2.$(SUFFIX) : spr2.c
2290+
$(CC) $(CFLAGS) -c -UCOMPLEX -UDOUBLE -o $(@F) $^
2291+
2292+
dspr2.$(SUFFIX) : spr2.c
2293+
$(CC) $(CFLAGS) -c -UCOMPLEX -DDOUBLE -o $(@F) $^
22002294

22012295
ssyrk.$(SUFFIX) : syrk.c
22022296
$(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)