Skip to content

Commit ce33da4

Browse files
authored
Merge pull request #2513 from aaawuanjun/develop
[OpenBlas]: Add benchmark tpsv file and modify benchmark/Makefile
2 parents a1b181c + 2428dc9 commit ce33da4

2 files changed

Lines changed: 260 additions & 0 deletions

File tree

benchmark/Makefile

Lines changed: 88 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -82,6 +82,7 @@ goto :: slinpack.goto dlinpack.goto clinpack.goto zlinpack.goto \
8282
sspmv.goto dspmv.goto \
8383
strmv.goto dtrmv.goto ctrmv.goto ztrmv.goto \
8484
stpmv.goto dtpmv.goto ctpmv.goto ztpmv.goto \
85+
stpsv.goto dtpsv.goto ctpsv.goto ztpsv.goto \
8586
strsv.goto dtrsv.goto ctrsv.goto ztrsv.goto \
8687
sgeev.goto dgeev.goto cgeev.goto zgeev.goto \
8788
sgesv.goto dgesv.goto cgesv.goto zgesv.goto \
@@ -120,6 +121,7 @@ acml :: slinpack.acml dlinpack.acml clinpack.acml zlinpack.acml \
120121
sgemv.acml dgemv.acml cgemv.acml zgemv.acml \
121122
strmv.acml dtrmv.acml ctrmv.acml ztrmv.acml \
122123
stpmv.acml dtpmv.acml ctpmv.acml ztpmv.acml \
124+
stpsv.acml dtpsv.acml ctpsv.acml ztpsv.acml \
123125
strsv.acml dtrsv.acml ctrsv.acml ztrsv.acml \
124126
sgeev.acml dgeev.acml cgeev.acml zgeev.acml \
125127
sgesv.acml dgesv.acml cgesv.acml zgesv.acml \
@@ -160,6 +162,7 @@ atlas :: slinpack.atlas dlinpack.atlas clinpack.atlas zlinpack.atlas \
160162
sspmv.atlas dspmv.atlas \
161163
strmv.atlas dtrmv.atlas ctrmv.atlas ztrmv.atlas \
162164
stpmv.atlas dtpmv.atlas ctpmv.atlas ztpmv.atlas \
165+
stpsv.atlas dtpsv.atlas ctpsv.atlas ztpsv.atlas \
163166
strsv.atlas dtrsv.atlas ctrsv.atlas ztrsv.atlas \
164167
sgeev.atlas dgeev.atlas cgeev.atlas zgeev.atlas \
165168
sgesv.atlas dgesv.atlas cgesv.atlas zgesv.atlas \
@@ -198,6 +201,7 @@ mkl :: slinpack.mkl dlinpack.mkl clinpack.mkl zlinpack.mkl \
198201
sgemv.mkl dgemv.mkl cgemv.mkl zgemv.mkl \
199202
strmv.mkl dtrmv.mkl ctrmv.mkl ztrmv.mkl \
200203
stpmv.mkl dtpmv.mkl ctpmv.mkl ztpmv.mkl \
204+
stpsv.mkl dtpsv.mkl ctpsv.mkl ztpsv.mkl \
201205
strsv.mkl dtrsv.mkl ctrsv.mkl ztrsv.mkl \
202206
sgeev.mkl dgeev.mkl cgeev.mkl zgeev.mkl \
203207
sgesv.mkl dgesv.mkl cgesv.mkl zgesv.mkl \
@@ -237,6 +241,7 @@ goto :: sgemm.goto dgemm.goto cgemm.goto zgemm.goto \
237241
sspmv.goto dspmv.goto \
238242
strmv.goto dtrmv.goto ctrmv.goto ztrmv.goto \
239243
stpmv.goto dtpmv.goto ctpmv.goto ztpmv.goto \
244+
stpsv.goto dtpsv.goto ctpsv.goto ztpsv.goto \
240245
strsv.goto dtrsv.goto ctrsv.goto ztrsv.goto \
241246
ssymm.goto dsymm.goto csymm.goto zsymm.goto \
242247
smallscaling \
@@ -281,6 +286,7 @@ acml :: slinpack.acml dlinpack.acml clinpack.acml zlinpack.acml \
281286
sgemv.acml dgemv.acml cgemv.acml zgemv.acml \
282287
strmv.acml dtrmv.acml ctrmv.acml ztrmv.acml \
283288
stpmv.acml dtpmv.acml ctpmv.acml ztpmv.acml \
289+
stpsv.acml dtpsv.acml ctpsv.acml ztpsv.acml \
284290
strsv.acml dtrsv.acml ctrsv.acml ztrsv.acml \
285291
sgeev.acml dgeev.acml cgeev.acml zgeev.acml \
286292
sgesv.acml dgesv.acml cgesv.acml zgesv.acml \
@@ -321,6 +327,7 @@ atlas :: slinpack.atlas dlinpack.atlas clinpack.atlas zlinpack.atlas \
321327
sspmv.atlas dspmv.atlas \
322328
strmv.atlas dtrmv.atlas ctrmv.atlas ztrmv.atlas \
323329
stpmv.atlas dtpmv.atlas ctpmv.atlas ztpmv.atlas \
330+
stpsv.atlas dtpsv.atlas ctpsv.atlas ztpsv.atlas \
324331
strsv.atlas dtrsv.atlas ctrsv.atlas ztrsv.atlas \
325332
sgeev.atlas dgeev.atlas cgeev.atlas zgeev.atlas \
326333
sgesv.atlas dgesv.atlas cgesv.atlas zgesv.atlas \
@@ -361,6 +368,7 @@ mkl :: slinpack.mkl dlinpack.mkl clinpack.mkl zlinpack.mkl \
361368
sgemv.mkl dgemv.mkl cgemv.mkl zgemv.mkl \
362369
strmv.mkl dtrmv.mkl ctrmv.mkl ztrmv.mkl \
363370
stpmv.mkl dtpmv.mkl ctpmv.mkl ztpmv.mkl \
371+
stpsv.mkl dtpsv.mkl ctpsv.mkl ztpsv.mkl \
364372
strsv.mkl dtrsv.mkl ctrsv.mkl ztrsv.mkl \
365373
sgeev.mkl dgeev.mkl cgeev.mkl zgeev.mkl \
366374
sgesv.mkl dgesv.mkl cgesv.mkl zgesv.mkl \
@@ -410,6 +418,7 @@ veclib :: slinpack.veclib dlinpack.veclib clinpack.veclib zlinpack.veclib \
410418
sgemv.veclib dgemv.veclib cgemv.veclib zgemv.veclib \
411419
strmv.veclib dtrmv.veclib ctrmv.veclib ztrmv.veclib \
412420
stpmv.veclib dtpmv.veclib ctpmv.veclib ztpmv.veclib \
421+
stpsv.veclib dtpsv.veclib ctpsv.veclib ztpsv.veclib \
413422
strsv.veclib dtrsv.veclib ctrsv.veclib ztrsv.veclib \
414423
sgeev.veclib dgeev.veclib cgeev.veclib zgeev.veclib \
415424
sgesv.veclib dgesv.veclib cgesv.veclib zgesv.veclib \
@@ -1456,6 +1465,73 @@ ztpmv.mkl : ztpmv.$(SUFFIX)
14561465

14571466
ztpmv.veclib : ztpmv.$(SUFFIX)
14581467
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1468+
1469+
##################################### Stpsv ####################################################
1470+
stpsv.goto : stpsv.$(SUFFIX) ../$(LIBNAME)
1471+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
1472+
1473+
stpsv.acml : stpsv.$(SUFFIX)
1474+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1475+
1476+
stpsv.atlas : stpsv.$(SUFFIX)
1477+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1478+
1479+
stpsv.mkl : stpsv.$(SUFFIX)
1480+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1481+
1482+
stpsv.veclib : stpsv.$(SUFFIX)
1483+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1484+
1485+
##################################### Dtpsv ####################################################
1486+
dtpsv.goto : dtpsv.$(SUFFIX) ../$(LIBNAME)
1487+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
1488+
1489+
dtpsv.acml : dtpsv.$(SUFFIX)
1490+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1491+
1492+
dtpsv.atlas : dtpsv.$(SUFFIX)
1493+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1494+
1495+
dtpsv.mkl : dtpsv.$(SUFFIX)
1496+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1497+
1498+
dtpsv.veclib : dtpsv.$(SUFFIX)
1499+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1500+
1501+
##################################### Ctpsv ####################################################
1502+
1503+
ctpsv.goto : ctpsv.$(SUFFIX) ../$(LIBNAME)
1504+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
1505+
1506+
ctpsv.acml : ctpsv.$(SUFFIX)
1507+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1508+
1509+
ctpsv.atlas : ctpsv.$(SUFFIX)
1510+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1511+
1512+
ctpsv.mkl : ctpsv.$(SUFFIX)
1513+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1514+
1515+
ctpsv.veclib : ctpsv.$(SUFFIX)
1516+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1517+
1518+
##################################### Ztpsv ####################################################
1519+
1520+
ztpsv.goto : ztpsv.$(SUFFIX) ../$(LIBNAME)
1521+
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
1522+
1523+
ztpsv.acml : ztpsv.$(SUFFIX)
1524+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBACML) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1525+
1526+
ztpsv.atlas : ztpsv.$(SUFFIX)
1527+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBATLAS) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1528+
1529+
ztpsv.mkl : ztpsv.$(SUFFIX)
1530+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBMKL) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1531+
1532+
ztpsv.veclib : ztpsv.$(SUFFIX)
1533+
-$(CC) $(CFLAGS) -o $(@F) $^ $(LIBVECLIB) $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB)
1534+
14591535
##################################### Strsv ####################################################
14601536
strsv.goto : strsv.$(SUFFIX) ../$(LIBNAME)
14611537
$(CC) $(CFLAGS) -o $(@F) $^ $(CEXTRALIB) $(EXTRALIB) $(FEXTRALIB) -lm
@@ -2910,6 +2986,18 @@ ctpmv.$(SUFFIX) : tpmv.c
29102986
ztpmv.$(SUFFIX) : tpmv.c
29112987
$(CC) $(CFLAGS) -c -DCOMPLEX -DDOUBLE -o $(@F) $^
29122988

2989+
stpsv.$(SUFFIX) : tpsv.c
2990+
$(CC) $(CFLAGS) -c -UCOMPLEX -UDOUBLE -o $(@F) $^
2991+
2992+
dtpsv.$(SUFFIX) : tpsv.c
2993+
$(CC) $(CFLAGS) -c -UCOMPLEX -DDOUBLE -o $(@F) $^
2994+
2995+
ctpsv.$(SUFFIX) : tpsv.c
2996+
$(CC) $(CFLAGS) -c -DCOMPLEX -UDOUBLE -o $(@F) $^
2997+
2998+
ztpsv.$(SUFFIX) : tpsv.c
2999+
$(CC) $(CFLAGS) -c -DCOMPLEX -DDOUBLE -o $(@F) $^
3000+
29133001
strsv.$(SUFFIX) : trsv.c
29143002
$(CC) $(CFLAGS) -c -UCOMPLEX -UDOUBLE -o $(@F) $^
29153003

benchmark/tpsv.c

Lines changed: 172 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,172 @@
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+
#undef TPSV
36+
37+
#ifndef COMPLEX
38+
39+
#ifdef DOUBLE
40+
#define TPSV BLASFUNC(dtpsv)
41+
#else
42+
#define TPSV BLASFUNC(stpsv)
43+
#endif
44+
45+
#else
46+
47+
#ifdef DOUBLE
48+
#define TPSV BLASFUNC(ztpsv)
49+
#else
50+
#define TPSV BLASFUNC(ctpsv)
51+
#endif
52+
53+
#endif
54+
55+
#if !defined(__WIN32__) && !defined(__WIN64__) && !defined(__CYGWIN32__) && 0
56+
57+
static void *huge_malloc(BLASLONG size)
58+
{
59+
int shmid;
60+
void *address;
61+
62+
#ifndef SHM_HUGETLB
63+
#define SHM_HUGETLB 04000
64+
#endif
65+
66+
if ((shmid =shmget(IPC_PRIVATE,
67+
(size + HUGE_PAGESIZE) & ~(HUGE_PAGESIZE - 1),
68+
SHM_HUGETLB | IPC_CREAT |0600)) < 0) {
69+
printf( "Memory allocation failed(shmget).\n");
70+
exit(1);
71+
}
72+
73+
address = shmat(shmid, NULL, SHM_RND);
74+
75+
if ((BLASLONG)address == -1) {
76+
printf( "Memory allocation failed(shmat).\n");
77+
exit(1);
78+
}
79+
80+
shmctl(shmid, IPC_RMID, 0);
81+
82+
return address;
83+
}
84+
85+
#define malloc huge_malloc
86+
87+
#endif
88+
89+
int main(int argc, char *argv[])
90+
{
91+
92+
FLOAT *a, *x;
93+
char *p;
94+
95+
char uplo ='U';
96+
char trans='N';
97+
char diag ='U';
98+
99+
int loops = 1;
100+
int l;
101+
blasint inc_x=1;
102+
103+
if ((p = getenv("OPENBLAS_UPLO"))) uplo=*p;
104+
if ((p = getenv("OPENBLAS_TRANS"))) trans=*p;
105+
if ((p = getenv("OPENBLAS_DIAG"))) diag=*p;
106+
if ((p = getenv("OPENBLAS_LOOPS"))) loops = atoi(p);
107+
if ((p = getenv("OPENBLAS_INCX"))) inc_x = atoi(p);
108+
109+
blasint n, i, j;
110+
111+
int from = 1;
112+
int to = 200;
113+
int step = 1;
114+
115+
struct timespec start = { 0, 0 }, stop = { 0, 0 };
116+
double time1, timeg;
117+
118+
argc--;argv++;
119+
120+
if (argc > 0) { from = atol(*argv); argc--; argv++;}
121+
if (argc > 0) { to = MAX(atol(*argv), from); argc--; argv++;}
122+
if (argc > 0) { step = atol(*argv); argc--; argv++;}
123+
124+
fprintf(stderr, "From : %3d To : %3d Step = %3d Uplo = %c Trans = %c Diag = %c Loops=%d Inc_x=%d\n", from,
125+
to, step, uplo, trans, diag, loops, inc_x);
126+
127+
if (( a = (FLOAT *)malloc(sizeof(FLOAT) * to * to * COMPSIZE)) == NULL) {
128+
fprintf(stderr,"Out of Memory!!\n");exit(1);
129+
}
130+
131+
if (( x = (FLOAT *)malloc(sizeof(FLOAT) * to * abs(inc_x) * COMPSIZE)) == NULL){
132+
fprintf(stderr,"Out of Memory!!\n");exit(1);
133+
}
134+
135+
#ifdef linux
136+
srandom(getpid());
137+
#endif
138+
139+
fprintf(stderr, " SIZE Flops\n");
140+
141+
for(n = from; n <= to; n += step) {
142+
timeg=0;
143+
144+
fprintf(stderr, " %6d : ", (int)n);
145+
for(j = 0; j < n; j++) {
146+
for(i = 0; i < n * COMPSIZE; i++) {
147+
a[(long)i + (long)j * (long)n * COMPSIZE] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
148+
}
149+
}
150+
151+
for (i = 0; i < n * COMPSIZE * abs(inc_x); i++) {
152+
x[i] = ((FLOAT) rand() / (FLOAT) RAND_MAX) - 0.5;
153+
}
154+
155+
for (l = 0; l < loops; l++) {
156+
clock_gettime(CLOCK_REALTIME, &start);
157+
TPSV (&uplo, &trans, &diag, &n, a, x, &inc_x);
158+
clock_gettime(CLOCK_REALTIME, &stop);
159+
160+
time1 = (double)(stop.tv_sec - start.tv_sec) + (double)((stop.tv_nsec - start.tv_nsec)) / 1.e9;
161+
timeg += time1;
162+
}
163+
164+
timeg /= loops;
165+
fprintf(stderr, " %10.2f MFlops %12.9f sec\n",
166+
COMPSIZE * COMPSIZE * 1. * (double)n * (double)n / timeg / 1.e6, timeg);
167+
}
168+
169+
return 0;
170+
}
171+
172+
// void main(int argc, char *argv[]) __attribute__((weak, alias("MAIN__")));

0 commit comments

Comments
 (0)