Initial Commit
This commit is contained in:
55
libs/libgrapheme-2.0.2/benchmark/case.c
Normal file
55
libs/libgrapheme-2.0.2/benchmark/case.c
Normal file
@@ -0,0 +1,55 @@
|
||||
/* See LICENSE file for copyright and license details. */
|
||||
#include <errno.h>
|
||||
#include <math.h>
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
|
||||
#include "../grapheme.h"
|
||||
#include "../gen/word-test.h"
|
||||
#include "util.h"
|
||||
|
||||
#define NUM_ITERATIONS 10000
|
||||
|
||||
struct break_benchmark_payload {
|
||||
uint_least32_t *src;
|
||||
size_t srclen;
|
||||
uint_least32_t *dest;
|
||||
size_t destlen;
|
||||
};
|
||||
|
||||
void
|
||||
libgrapheme(const void *payload)
|
||||
{
|
||||
const struct break_benchmark_payload *p = payload;
|
||||
|
||||
grapheme_to_uppercase(p->src, p->srclen, p->dest, p->destlen);
|
||||
}
|
||||
|
||||
int
|
||||
main(int argc, char *argv[])
|
||||
{
|
||||
struct break_benchmark_payload p;
|
||||
double baseline = (double)NAN;
|
||||
|
||||
(void)argc;
|
||||
|
||||
if ((p.src = generate_cp_test_buffer(word_break_test,
|
||||
LEN(word_break_test),
|
||||
&(p.srclen))) == NULL) {
|
||||
return 1;
|
||||
}
|
||||
if ((p.dest = calloc((p.destlen = 2 * p.srclen), sizeof(*(p.dest)))) == NULL) {
|
||||
fprintf(stderr, "calloc: Out of memory\n");
|
||||
}
|
||||
|
||||
printf("%s\n", argv[0]);
|
||||
run_benchmark(libgrapheme, &p, "libgrapheme ", NULL, "codepoint",
|
||||
&baseline, NUM_ITERATIONS, p.srclen - 1);
|
||||
|
||||
free(p.src);
|
||||
free(p.dest);
|
||||
|
||||
return 0;
|
||||
}
|
||||
86
libs/libgrapheme-2.0.2/benchmark/character.c
Normal file
86
libs/libgrapheme-2.0.2/benchmark/character.c
Normal file
@@ -0,0 +1,86 @@
|
||||
/* See LICENSE file for copyright and license details. */
|
||||
#include <errno.h>
|
||||
#include <math.h>
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
|
||||
#include "../grapheme.h"
|
||||
#include "../gen/character-test.h"
|
||||
#include "util.h"
|
||||
|
||||
#include <utf8proc.h>
|
||||
|
||||
#define NUM_ITERATIONS 100000
|
||||
|
||||
struct break_benchmark_payload {
|
||||
uint_least32_t *buf;
|
||||
utf8proc_int32_t *buf_utf8proc;
|
||||
size_t buflen;
|
||||
};
|
||||
|
||||
void
|
||||
libgrapheme(const void *payload)
|
||||
{
|
||||
uint_least16_t state = 0;
|
||||
const struct break_benchmark_payload *p = payload;
|
||||
size_t i;
|
||||
|
||||
for (i = 0; i + 1 < p->buflen; i++) {
|
||||
(void)grapheme_is_character_break(p->buf[i], p->buf[i+1],
|
||||
&state);
|
||||
}
|
||||
}
|
||||
|
||||
void
|
||||
libutf8proc(const void *payload)
|
||||
{
|
||||
utf8proc_int32_t state = 0;
|
||||
const struct break_benchmark_payload *p = payload;
|
||||
size_t i;
|
||||
|
||||
for (i = 0; i + 1 < p->buflen; i++) {
|
||||
(void)utf8proc_grapheme_break_stateful(p->buf_utf8proc[i],
|
||||
p->buf_utf8proc[i+1],
|
||||
&state);
|
||||
}
|
||||
}
|
||||
|
||||
int
|
||||
main(int argc, char *argv[])
|
||||
{
|
||||
struct break_benchmark_payload p;
|
||||
double baseline = (double)NAN;
|
||||
size_t i;
|
||||
|
||||
(void)argc;
|
||||
|
||||
if ((p.buf = generate_cp_test_buffer(character_break_test,
|
||||
LEN(character_break_test),
|
||||
&(p.buflen))) == NULL) {
|
||||
return 1;
|
||||
}
|
||||
if ((p.buf_utf8proc = malloc(p.buflen * sizeof(*(p.buf_utf8proc)))) == NULL) {
|
||||
fprintf(stderr, "malloc: %s\n", strerror(errno));
|
||||
exit(1);
|
||||
}
|
||||
for (i = 0; i < p.buflen; i++) {
|
||||
/*
|
||||
* there is no overflow, as we know that the maximum
|
||||
* codepoint is 0x10FFFF, which is way below 2^31
|
||||
*/
|
||||
p.buf_utf8proc[i] = (utf8proc_int32_t)p.buf[i];
|
||||
}
|
||||
|
||||
printf("%s\n", argv[0]);
|
||||
run_benchmark(libgrapheme, &p, "libgrapheme ", NULL, "comparison",
|
||||
&baseline, NUM_ITERATIONS, p.buflen - 1);
|
||||
run_benchmark(libutf8proc, &p, "libutf8proc ", NULL, "comparison",
|
||||
&baseline, NUM_ITERATIONS, p.buflen - 1);
|
||||
|
||||
free(p.buf);
|
||||
free(p.buf_utf8proc);
|
||||
|
||||
return 0;
|
||||
}
|
||||
52
libs/libgrapheme-2.0.2/benchmark/line.c
Normal file
52
libs/libgrapheme-2.0.2/benchmark/line.c
Normal file
@@ -0,0 +1,52 @@
|
||||
/* See LICENSE file for copyright and license details. */
|
||||
#include <errno.h>
|
||||
#include <math.h>
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
|
||||
#include "../grapheme.h"
|
||||
#include "../gen/line-test.h"
|
||||
#include "util.h"
|
||||
|
||||
#define NUM_ITERATIONS 10000
|
||||
|
||||
struct break_benchmark_payload {
|
||||
uint_least32_t *buf;
|
||||
size_t buflen;
|
||||
};
|
||||
|
||||
void
|
||||
libgrapheme(const void *payload)
|
||||
{
|
||||
const struct break_benchmark_payload *p = payload;
|
||||
size_t off;
|
||||
|
||||
for (off = 0; off < p->buflen; ) {
|
||||
off += grapheme_next_line_break(p->buf + off, p->buflen - off);
|
||||
}
|
||||
}
|
||||
|
||||
int
|
||||
main(int argc, char *argv[])
|
||||
{
|
||||
struct break_benchmark_payload p;
|
||||
double baseline = (double)NAN;
|
||||
|
||||
(void)argc;
|
||||
|
||||
if ((p.buf = generate_cp_test_buffer(line_break_test,
|
||||
LEN(line_break_test),
|
||||
&(p.buflen))) == NULL) {
|
||||
return 1;
|
||||
}
|
||||
|
||||
printf("%s\n", argv[0]);
|
||||
run_benchmark(libgrapheme, &p, "libgrapheme ", NULL, "codepoint",
|
||||
&baseline, NUM_ITERATIONS, p.buflen - 1);
|
||||
|
||||
free(p.buf);
|
||||
|
||||
return 0;
|
||||
}
|
||||
52
libs/libgrapheme-2.0.2/benchmark/sentence.c
Normal file
52
libs/libgrapheme-2.0.2/benchmark/sentence.c
Normal file
@@ -0,0 +1,52 @@
|
||||
/* See LICENSE file for copyright and license details. */
|
||||
#include <errno.h>
|
||||
#include <math.h>
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
|
||||
#include "../grapheme.h"
|
||||
#include "../gen/sentence-test.h"
|
||||
#include "util.h"
|
||||
|
||||
#define NUM_ITERATIONS 100000
|
||||
|
||||
struct break_benchmark_payload {
|
||||
uint_least32_t *buf;
|
||||
size_t buflen;
|
||||
};
|
||||
|
||||
void
|
||||
libgrapheme(const void *payload)
|
||||
{
|
||||
const struct break_benchmark_payload *p = payload;
|
||||
size_t off;
|
||||
|
||||
for (off = 0; off < p->buflen; ) {
|
||||
off += grapheme_next_sentence_break(p->buf + off, p->buflen - off);
|
||||
}
|
||||
}
|
||||
|
||||
int
|
||||
main(int argc, char *argv[])
|
||||
{
|
||||
struct break_benchmark_payload p;
|
||||
double baseline = (double)NAN;
|
||||
|
||||
(void)argc;
|
||||
|
||||
if ((p.buf = generate_cp_test_buffer(sentence_break_test,
|
||||
LEN(sentence_break_test),
|
||||
&(p.buflen))) == NULL) {
|
||||
return 1;
|
||||
}
|
||||
|
||||
printf("%s\n", argv[0]);
|
||||
run_benchmark(libgrapheme, &p, "libgrapheme ", NULL, "codepoint",
|
||||
&baseline, NUM_ITERATIONS, p.buflen - 1);
|
||||
|
||||
free(p.buf);
|
||||
|
||||
return 0;
|
||||
}
|
||||
95
libs/libgrapheme-2.0.2/benchmark/utf8-decode.c
Normal file
95
libs/libgrapheme-2.0.2/benchmark/utf8-decode.c
Normal file
@@ -0,0 +1,95 @@
|
||||
/* See LICENSE file for copyright and license details. */
|
||||
#include <errno.h>
|
||||
#include <math.h>
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
|
||||
#include "../grapheme.h"
|
||||
#include "../gen/character-test.h"
|
||||
#include "util.h"
|
||||
|
||||
#include <utf8proc.h>
|
||||
|
||||
#define NUM_ITERATIONS 100000
|
||||
|
||||
struct utf8_benchmark_payload {
|
||||
char *buf;
|
||||
utf8proc_uint8_t *buf_utf8proc;
|
||||
size_t buflen;
|
||||
};
|
||||
|
||||
void
|
||||
libgrapheme(const void *payload)
|
||||
{
|
||||
const struct utf8_benchmark_payload *p = payload;
|
||||
uint_least32_t cp;
|
||||
size_t ret, off;
|
||||
|
||||
for (off = 0; off < p->buflen; off += ret) {
|
||||
if ((ret = grapheme_decode_utf8(p->buf + off,
|
||||
p->buflen - off, &cp)) >
|
||||
(p->buflen - off)) {
|
||||
break;
|
||||
}
|
||||
(void)cp;
|
||||
}
|
||||
}
|
||||
|
||||
void
|
||||
libutf8proc(const void *payload)
|
||||
{
|
||||
const struct utf8_benchmark_payload *p = payload;
|
||||
utf8proc_int32_t cp;
|
||||
utf8proc_ssize_t ret;
|
||||
size_t off;
|
||||
|
||||
for (off = 0; off < p->buflen; off += (size_t)ret) {
|
||||
if ((ret = utf8proc_iterate(p->buf_utf8proc + off,
|
||||
(utf8proc_ssize_t)(p->buflen - off),
|
||||
&cp)) < 0) {
|
||||
break;
|
||||
}
|
||||
(void)cp;
|
||||
}
|
||||
}
|
||||
|
||||
int
|
||||
main(int argc, char *argv[])
|
||||
{
|
||||
struct utf8_benchmark_payload p;
|
||||
size_t i;
|
||||
double baseline = (double)NAN;
|
||||
|
||||
(void)argc;
|
||||
|
||||
p.buf = generate_utf8_test_buffer(character_break_test,
|
||||
LEN(character_break_test),
|
||||
&(p.buflen));
|
||||
|
||||
/* convert cp-buffer to stupid custom libutf8proc-uint8-type */
|
||||
if ((p.buf_utf8proc = malloc(p.buflen)) == NULL) {
|
||||
fprintf(stderr, "malloc: %s\n", strerror(errno));
|
||||
exit(1);
|
||||
}
|
||||
for (i = 0; i < p.buflen; i++) {
|
||||
/*
|
||||
* even if char is larger than 8 bit, it will only have
|
||||
* any of the first 8 bits set (by construction).
|
||||
*/
|
||||
p.buf_utf8proc[i] = (utf8proc_uint8_t)p.buf[i];
|
||||
}
|
||||
|
||||
printf("%s\n", argv[0]);
|
||||
run_benchmark(libgrapheme, &p, "libgrapheme ", NULL,
|
||||
"byte", &baseline, NUM_ITERATIONS, p.buflen);
|
||||
run_benchmark(libutf8proc, &p, "libutf8proc ",
|
||||
"but unsafe (does not detect overlong encodings)",
|
||||
"byte", &baseline, NUM_ITERATIONS, p.buflen);
|
||||
|
||||
free(p.buf);
|
||||
free(p.buf_utf8proc);
|
||||
|
||||
return 0;
|
||||
}
|
||||
115
libs/libgrapheme-2.0.2/benchmark/util.c
Normal file
115
libs/libgrapheme-2.0.2/benchmark/util.c
Normal file
@@ -0,0 +1,115 @@
|
||||
/* See LICENSE file for copyright and license details. */
|
||||
#include <math.h>
|
||||
#include <stdlib.h>
|
||||
#include <stdio.h>
|
||||
#include <time.h>
|
||||
|
||||
#include "../gen/types.h"
|
||||
#include "../grapheme.h"
|
||||
#include "util.h"
|
||||
|
||||
uint_least32_t *
|
||||
generate_cp_test_buffer(const struct break_test *test, size_t testlen,
|
||||
size_t *buflen)
|
||||
{
|
||||
size_t i, j, off;
|
||||
uint_least32_t *buf;
|
||||
|
||||
/* allocate and generate buffer */
|
||||
for (i = 0, *buflen = 0; i < testlen; i++) {
|
||||
*buflen += test[i].cplen;
|
||||
}
|
||||
if (!(buf = calloc(*buflen, sizeof(*buf)))) {
|
||||
fprintf(stderr, "generate_test_buffer: calloc: Out of memory.\n");
|
||||
exit(1);
|
||||
}
|
||||
for (i = 0, off = 0; i < testlen; i++) {
|
||||
for (j = 0; j < test[i].cplen; j++) {
|
||||
buf[off + j] = test[i].cp[j];
|
||||
}
|
||||
off += test[i].cplen;
|
||||
}
|
||||
|
||||
return buf;
|
||||
}
|
||||
|
||||
char *
|
||||
generate_utf8_test_buffer(const struct break_test *test, size_t testlen,
|
||||
size_t *buflen)
|
||||
{
|
||||
size_t i, j, off, ret;
|
||||
char *buf;
|
||||
|
||||
/* allocate and generate buffer */
|
||||
for (i = 0, *buflen = 0; i < testlen; i++) {
|
||||
for (j = 0; j < test[i].cplen; j++) {
|
||||
*buflen += grapheme_encode_utf8(test[i].cp[j], NULL, 0);
|
||||
}
|
||||
}
|
||||
(*buflen)++; /* terminating NUL-byte */
|
||||
if (!(buf = malloc(*buflen))) {
|
||||
fprintf(stderr, "generate_test_buffer: malloc: Out of memory.\n");
|
||||
exit(1);
|
||||
}
|
||||
for (i = 0, off = 0; i < testlen; i++) {
|
||||
for (j = 0; j < test[i].cplen; j++, off += ret) {
|
||||
if ((ret = grapheme_encode_utf8(test[i].cp[j],
|
||||
buf + off,
|
||||
*buflen - off)) >
|
||||
(*buflen - off)) {
|
||||
/* shouldn't happen */
|
||||
fprintf(stderr, "generate_utf8_test_buffer: "
|
||||
"Buffer too small.\n");
|
||||
exit(1);
|
||||
}
|
||||
}
|
||||
}
|
||||
buf[*buflen - 1] = '\0';
|
||||
|
||||
return buf;
|
||||
}
|
||||
|
||||
static double
|
||||
time_diff(struct timespec *a, struct timespec *b)
|
||||
{
|
||||
return (double)(b->tv_sec - a->tv_sec) +
|
||||
(double)(b->tv_nsec - a->tv_nsec) * 1E-9;
|
||||
}
|
||||
|
||||
void
|
||||
run_benchmark(void (*func)(const void *), const void *payload,
|
||||
const char *name, const char *comment, const char *unit,
|
||||
double *baseline, size_t num_iterations,
|
||||
size_t units_per_iteration)
|
||||
{
|
||||
struct timespec start, end;
|
||||
size_t i;
|
||||
double diff;
|
||||
|
||||
printf("\t%s ", name);
|
||||
fflush(stdout);
|
||||
|
||||
clock_gettime(CLOCK_MONOTONIC, &start);
|
||||
for (i = 0; i < num_iterations; i++) {
|
||||
func(payload);
|
||||
|
||||
if (i % (num_iterations / 10) == 0) {
|
||||
printf(".");
|
||||
fflush(stdout);
|
||||
}
|
||||
}
|
||||
clock_gettime(CLOCK_MONOTONIC, &end);
|
||||
diff = time_diff(&start, &end) / (double)num_iterations /
|
||||
(double)units_per_iteration;
|
||||
|
||||
if (isnan(*baseline)) {
|
||||
*baseline = diff;
|
||||
printf(" avg. %.3es/%s (baseline)\n", diff, unit);
|
||||
} else {
|
||||
printf(" avg. %.3es/%s (%.2f%% %s%s%s)\n", diff, unit,
|
||||
fabs(1.0 - diff / *baseline) * 100,
|
||||
(diff < *baseline) ? "faster" : "slower",
|
||||
comment ? ", " : "",
|
||||
comment ? comment : "");
|
||||
}
|
||||
}
|
||||
23
libs/libgrapheme-2.0.2/benchmark/util.h
Normal file
23
libs/libgrapheme-2.0.2/benchmark/util.h
Normal file
@@ -0,0 +1,23 @@
|
||||
/* See LICENSE file for copyright and license details. */
|
||||
#ifndef UTIL_H
|
||||
#define UTIL_H
|
||||
|
||||
#include "../gen/types.h"
|
||||
|
||||
#define LEN(x) (sizeof(x) / sizeof(*(x)))
|
||||
|
||||
#ifdef __has_attribute
|
||||
#if __has_attribute(optnone)
|
||||
void libgrapheme(const void *) __attribute__((optnone));
|
||||
void libutf8proc(const void *) __attribute__((optnone));
|
||||
#endif
|
||||
#endif
|
||||
|
||||
uint_least32_t *generate_cp_test_buffer(const struct break_test *, size_t,
|
||||
size_t *);
|
||||
char *generate_utf8_test_buffer(const struct break_test *, size_t, size_t *);
|
||||
|
||||
void run_benchmark(void (*func)(const void *), const void *, const char *,
|
||||
const char *, const char *, double *, size_t, size_t);
|
||||
|
||||
#endif /* UTIL_H */
|
||||
52
libs/libgrapheme-2.0.2/benchmark/word.c
Normal file
52
libs/libgrapheme-2.0.2/benchmark/word.c
Normal file
@@ -0,0 +1,52 @@
|
||||
/* See LICENSE file for copyright and license details. */
|
||||
#include <errno.h>
|
||||
#include <math.h>
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
|
||||
#include "../grapheme.h"
|
||||
#include "../gen/word-test.h"
|
||||
#include "util.h"
|
||||
|
||||
#define NUM_ITERATIONS 10000
|
||||
|
||||
struct break_benchmark_payload {
|
||||
uint_least32_t *buf;
|
||||
size_t buflen;
|
||||
};
|
||||
|
||||
void
|
||||
libgrapheme(const void *payload)
|
||||
{
|
||||
const struct break_benchmark_payload *p = payload;
|
||||
size_t off;
|
||||
|
||||
for (off = 0; off < p->buflen; ) {
|
||||
off += grapheme_next_word_break(p->buf + off, p->buflen - off);
|
||||
}
|
||||
}
|
||||
|
||||
int
|
||||
main(int argc, char *argv[])
|
||||
{
|
||||
struct break_benchmark_payload p;
|
||||
double baseline = (double)NAN;
|
||||
|
||||
(void)argc;
|
||||
|
||||
if ((p.buf = generate_cp_test_buffer(word_break_test,
|
||||
LEN(word_break_test),
|
||||
&(p.buflen))) == NULL) {
|
||||
return 1;
|
||||
}
|
||||
|
||||
printf("%s\n", argv[0]);
|
||||
run_benchmark(libgrapheme, &p, "libgrapheme ", NULL, "codepoint",
|
||||
&baseline, NUM_ITERATIONS, p.buflen - 1);
|
||||
|
||||
free(p.buf);
|
||||
|
||||
return 0;
|
||||
}
|
||||
Reference in New Issue
Block a user