mirror of
https://github.com/opencv/opencv.git
synced 2024-12-16 10:29:11 +08:00
0de26fd78e
Zlib-ng is zlib replacement with optimizations for "next generation" systems. Its optimization may benifits image library decode and encode speed such as libpng. In our tests, if using zlib-ng and libpng combination on a x86_64 machine with AVX2, the time of `imdecode` amd `imencode` will drop 20% approximately. This patch enables zlib-ng's optimization if `CV_DISABLE_OPTIMIZATION` is OFF. Since Zlib-ng can dispatch intrinsics on the fly, port work is much easier. Related discussion: https://github.com/opencv/opencv/issues/22573
32 lines
828 B
C
32 lines
828 B
C
/* Optimized slide_hash for PowerPC processors
|
|
* Copyright (C) 2017-2021 Mika T. Lindqvist <postmaster@raasu.org>
|
|
* For conditions of distribution and use, see copyright notice in zlib.h
|
|
*/
|
|
|
|
#include <altivec.h>
|
|
#include "zbuild.h"
|
|
#include "deflate.h"
|
|
|
|
static inline void slide_hash_chain(Pos *table, uint32_t entries, uint16_t wsize) {
|
|
const vector unsigned short vmx_wsize = vec_splats(wsize);
|
|
Pos *p = table;
|
|
|
|
do {
|
|
vector unsigned short value, result;
|
|
|
|
value = vec_ld(0, p);
|
|
result = vec_subs(value, vmx_wsize);
|
|
vec_st(result, 0, p);
|
|
|
|
p += 8;
|
|
entries -= 8;
|
|
} while (entries > 0);
|
|
}
|
|
|
|
void Z_INTERNAL SLIDE_PPC(deflate_state *s) {
|
|
uint16_t wsize = s->w_size;
|
|
|
|
slide_hash_chain(s->head, HASH_SIZE, wsize);
|
|
slide_hash_chain(s->prev, wsize, wsize);
|
|
}
|