1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383
|
/* Copyright (C) 2001-2021 Artifex Software, Inc.
All Rights Reserved.
This software is provided AS-IS with no warranty, either express or
implied.
This software is distributed under license and may not be copied,
modified or distributed except as expressly authorized under the terms
of the license contained in the file LICENSE in this distribution.
Refer to licensing information at http://www.artifex.com or contact
Artifex Software, Inc., 1305 Grant Avenue - Suite 200, Novato,
CA 94945, U.S.A., +1(415)492-9861, for further information.
*/
/* Simple hardware benchmarking suite (C and PostScript) */
#include "stdio_.h"
#include <stdlib.h>
/* Patchup for GS externals */
FILE *gs_stdout;
FILE *gs_stderr;
const char gp_scratch_file_name_prefix[] = "gs_";
static void
capture_stdio(void)
{
gs_stdout = stdout;
gs_stderr = stderr;
}
#include "gsio.h"
#undef gs_stdout
#undef gs_stderr
#undef stdout
#define stdout gs_stdout
#undef stderr
#define stderr gs_stderr
FILE *
gp_open_scratch_file(const gs_memory_t *mem,
const char *prefix,
char *fname,
const char *mode)
{
return NULL;
}
void
gp_set_printer_binary(int prnfno, int binary)
{
}
void
gs_to_exit(int n)
{
}
#define eprintf_program_ident(f, pn, rn) (void)0
void
lprintf_file_and_line(FILE * f, const char *file, int line)
{
fprintf(f, "%s(%d): ", file, line);
}
/*
* Read the CPU time (in seconds since an implementation-defined epoch)
* into ptm[0], and fraction (in nanoseconds) into ptm[1].
*/
#include "gp_unix.c"
#undef stdout
#define stdout gs_stdout
#undef stderr
#define stderr gs_stderr
/* Loop unrolling macros */
#define do10(x) x;x;x;x;x; x;x;x;x;x
/* Define the actual benchmarks. */
static int
iadd(int a, int n, char **msg)
{
int b = 0, i;
for (i = n / 20; --i >= 0;) {
do10((b += a, b += i));
}
*msg = "integer adds";
return b;
}
static int
imul(int a, int n, char **msg)
{
int b = 1, i;
for (i = n / 20; --i > 0;) {
do10((b *= a, b *= i));
}
*msg = "integer multiplies";
return b;
}
static int
idiv(int a, int n, char **msg)
{
int b = 1, i;
for (i = n / 20; --i > 0;) {
b += 999999;
do10((b /= a, b /= i));
}
*msg = "integer divides";
return b;
}
static int
fadd(float a, int n, char **msg)
{
float b = 0;
int i;
for (i = n / 10; --i >= 0;) {
do10((b += a));
}
*msg = "floating adds";
return b;
}
static int
fmul(float a, int n, char **msg)
{
float b = 1;
int i;
for (i = n / 10; --i >= 0;) {
do10((b *= a));
}
*msg = "floating multiplies";
return b;
}
static int
fdiv(float a, int n, char **msg)
{
float b = 1;
int i;
for (i = n / 10; --i >= 0;) {
do10((b /= a));
}
*msg = "floating divides";
return b;
}
static int
fconv(int a, int n, char **msg)
{
int b[10];
float f[10];
int i;
b[0] = a;
for (i = n / 20; --i >= 0;)
f[0] = b[0], f[1] = b[1], f[2] = b[2], f[3] = b[3], f[4] = b[4],
f[5] = b[5], f[6] = b[6], f[7] = b[7], f[8] = b[8], f[9] = b[9],
b[0] = f[1], b[1] = f[2], b[2] = f[3], b[3] = f[4], b[4] = f[5],
b[5] = f[6], b[6] = f[7], b[7] = f[8], b[8] = f[9], b[9] = f[0];
*msg = "float/int conversions";
return b[0];
}
static int
mfast(int *m, int n, char **msg)
{
int i;
m[0] = n;
for (i = n / 20; --i >= 0;)
m[9] = m[8], m[8] = m[7], m[7] = m[6], m[6] = m[5], m[5] = m[4],
m[4] = m[3], m[3] = m[2], m[2] = m[1], m[1] = m[0], m[0] = m[9];
*msg = "fast memory accesses";
return m[0];
}
static int
mslow(int *m, int n, char **msg)
{
int *p;
int i, k = 0;
m[0] = n;
for (i = n / 20; --i >= 0; k = (k + 397) & 0x3ffff)
p = m + k,
p[0] = p[100], p[20] = p[120], p[40] = p[140],
p[60] = p[160], p[80] = p[180],
p[200] = p[300], p[220] = p[320], p[240] = p[340],
p[260] = p[360], p[280] = p[380];
*msg = "slow memory accesses";
return m[0];
}
int
main(int argc, const char *argv[])
{
int i;
int *mem = malloc(1100000);
capture_stdio();
for (i = 0;; ++i) {
long t0[2], t1[2];
char *msg;
int n;
gp_get_usertime(t0);
switch (i) {
case 0:
iadd(0, n = 10000000, &msg);
break;
case 1:
imul(1, n = 1000000, &msg);
break;
case 2:
idiv(1, n = 1000000, &msg);
break;
case 3:
fadd(3.14, n = 10000000, &msg);
break;
case 4:
fmul(1.0000001, n = 10000000, &msg);
break;
case 5:
fdiv(1.0000001, n = 1000000, &msg);
break;
case 6:
fconv(12345, n = 10000000, &msg);
break;
case 7:
mfast(mem, n = 10000000, &msg);
break;
case 8:
mslow(mem, n = 1000000, &msg);
break;
default:
free(mem);
exit(0);
}
gp_get_usertime(t1);
fprintf(stdout, "Time for %9d %s = %g ms\n", n, msg,
(t1[0] - t0[0]) * 1000.0 + (t1[1] - t0[1]) / 1000000.0);
fflush(stdout);
}
}
/*
Output from SPARCstation 10, gcc -O bench.c gp_unix.c:
Time for 10000000 integer adds = 113.502 ms
Time for 1000000 integer multiplies = 467.965 ms
Time for 1000000 integer divides = 594.328 ms
Time for 10000000 floating adds = 641.21 ms
Time for 10000000 floating multiplies = 643.357 ms
Time for 1000000 floating divides = 131.995 ms
Time for 10000000 float/int conversions = 602.061 ms
Time for 10000000 fast memory accesses = 201.048 ms
Time for 1000000 slow memory accesses = 552.606 ms
Output from 486DX/25, wcl386 -oit bench.c gp_iwatc.c gp_msdos.c:
Time for 10000000 integer adds = 490 ms
Time for 1000000 integer multiplies = 770 ms
Time for 1000000 integer divides = 1860 ms
Time for 10000000 floating adds = 4070 ms
Time for 10000000 floating multiplies = 4450 ms
Time for 1000000 floating divides = 2470 ms
Time for 10000000 float/int conversions = 25650 ms
Time for 10000000 fast memory accesses = 990 ms
Time for 1000000 slow memory accesses = 330 ms
*/
/*
The rest of this file contains a similar benchmark in PostScript.
%!
/timer % <str> <N> <proc> timer
{ bind 2 copy usertime mark 4 2 roll repeat cleartomark usertime exch sub
% Stack: str N proc dt
exch pop
(Time for ) print exch =only ( ) print exch =only (: ) print
=only ( ms
) print flush
} def
(x 20 integer adds) 5000 { 0
0 add 0 add 0 add 0 add 0 add 0 add 0 add 0 add 0 add 0 add
0 add 0 add 0 add 0 add 0 add 0 add 0 add 0 add 0 add 0 add
pop } timer
(x 20 integer multiplies) 5000 { 1
3 mul 3 mul 3 mul 3 mul 3 mul 3 mul 3 mul 3 mul 3 mul 3 mul
3 mul 3 mul 3 mul 3 mul 3 mul 3 mul 3 mul 3 mul 3 mul 3 mul
pop } timer
(x 20 integer divides) 5000 { 1000000000
3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv
3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv 3 idiv
pop } timer
(x 20 floating adds) 5000 { 0.0
1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add
1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add 1.0 add
pop } timer
(x 20 floating multiplies) 5000 { 1.0
2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul
2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul 2.3 mul
pop } timer
(x 20 floating divides) 5000 { 1.0
2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div
2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div 2.3 div
pop } timer
(x 20 float/int conversions) 5000 { 12345.0
cvi cvr cvi cvr cvi cvr cvi cvr cvi cvr
cvi cvr cvi cvr cvi cvr cvi cvr cvi cvr
pop } timer
/S 2048 string def
(x 10000(byte) fast memory accesses) 1000 {
//S 1024 1000 getinterval //S copy pop
//S 1024 1000 getinterval //S copy pop
//S 1024 1000 getinterval //S copy pop
//S 1024 1000 getinterval //S copy pop
//S 1024 1000 getinterval //S copy pop
} timer
/A [ 500 { 2048 string } repeat ] def
(x 500 x 2000(byte) slower memory accesses) 10 {
0 1 499 {
//A exch get dup 1024 1000 getinterval exch copy pop
} for
} timer
/Times-Roman findfont 36 scalefont setfont
currentcacheparams pop pop 0 1 index setcacheparams
/D [4 0 0 4 0 0] 1440 1440 <00 ff> makeimagedevice def
D setdevice
72 72 translate
gsave 15 rotate
0 0 moveto (A) show
(x 10 (A) show (cache)) 100 {
0 0 moveto
(A) show (A) show (A) show (A) show (A) show
(A) show (A) show (A) show (A) show (A) show
} timer grestore
0 setcachelimit
gsave 10 rotate
(x 10 (A) show (no cache)) 10 {
0 0 moveto
(A) show (A) show (A) show (A) show (A) show
(A) show (A) show (A) show (A) show (A) show
} timer grestore
quit
Results for SUN Sparc 2 (rated at 25 MIPS according to manual)
./gs
Now in gs_init.ps
TextAlphaBits defined GraphicsAlphaBits defined
Aladdin Ghostscript 3.50 (1995-9-24)
(c) 1995 Aladdin Enterprises, Menlo Park, CA. All rights reserved. This
software comes with NO WARRANTY: see the file PUBLIC for details. Leaving
gs_init.ps
GS>(ben1.c) run
Time for 5000 x 20 integer adds: 171 ms
Time for 5000 x 20 integer multiplies: 504 ms
Time for 5000 x 20 integer divides: 334 ms
Time for 5000 x 20 floating adds: 148 ms
Time for 5000 x 20 floating multiplies: 165 ms
Time for 5000 x 20 floating divides: 194 ms
Time for 5000 x 20 float/int conversions: 121 ms
Time for 1000 x 10000(byte) fast memory accesses: 112 ms
Time for 10 x 500 x 2000(byte) slower memory accesses: 236 ms
Loading NimbusRomanNo9L-Regular font from
[...]/n021003l.gsf... 1739080 414724 2564864 1251073 0
done. Time for 100 x 10 (A) show (cache): 144 ms
Time for 10 x 10 (A) show (no cache): 538 ms
Output from SPARCstation 10, gs 3.60 compiled with gcc -g -O -DDEBUG:
gsnd bench.c
Aladdin Ghostscript 3.60 (1995-10-23)
|