mirror of
https://github.com/opencv/opencv.git
synced 2024-12-26 02:28:04 +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
53 lines
1.8 KiB
C
53 lines
1.8 KiB
C
/* slide_hash.c -- slide hash table C implementation
|
|
*
|
|
* Copyright (C) 1995-2013 Jean-loup Gailly and Mark Adler
|
|
* For conditions of distribution and use, see copyright notice in zlib.h
|
|
*/
|
|
|
|
#include "zbuild.h"
|
|
#include "deflate.h"
|
|
|
|
/* ===========================================================================
|
|
* Slide the hash table when sliding the window down (could be avoided with 32
|
|
* bit values at the expense of memory usage). We slide even when level == 0 to
|
|
* keep the hash table consistent if we switch back to level > 0 later.
|
|
*/
|
|
static inline void slide_hash_c_chain(Pos *table, uint32_t entries, uint16_t wsize) {
|
|
#ifdef NOT_TWEAK_COMPILER
|
|
table += entries;
|
|
do {
|
|
unsigned m;
|
|
m = *--table;
|
|
*table = (Pos)(m >= wsize ? m-wsize : 0);
|
|
/* If entries is not on any hash chain, prev[entries] is garbage but
|
|
* its value will never be used.
|
|
*/
|
|
} while (--entries);
|
|
#else
|
|
{
|
|
/* As of I make this change, gcc (4.8.*) isn't able to vectorize
|
|
* this hot loop using saturated-subtraction on x86-64 architecture.
|
|
* To avoid this defect, we can change the loop such that
|
|
* o. the pointer advance forward, and
|
|
* o. demote the variable 'm' to be local to the loop, and
|
|
* choose type "Pos" (instead of 'unsigned int') for the
|
|
* variable to avoid unnecessary zero-extension.
|
|
*/
|
|
unsigned int i;
|
|
Pos *q = table;
|
|
for (i = 0; i < entries; i++) {
|
|
Pos m = *q;
|
|
Pos t = (Pos)wsize;
|
|
*q++ = (Pos)(m >= t ? m-t: 0);
|
|
}
|
|
}
|
|
#endif /* NOT_TWEAK_COMPILER */
|
|
}
|
|
|
|
Z_INTERNAL void slide_hash_c(deflate_state *s) {
|
|
uint16_t wsize = (uint16_t)s->w_size;
|
|
|
|
slide_hash_c_chain(s->head, HASH_SIZE, wsize);
|
|
slide_hash_c_chain(s->prev, wsize, wsize);
|
|
}
|