mirror of
https://github.com/opencv/opencv.git
synced 2024-12-16 18:39:12 +08:00
85923c8f30
Update zlib-ng to 2.2.1 #26113 Release: https://github.com/zlib-ng/zlib-ng/releases/tag/2.2.1 ARM diagnostics patch: https://github.com/zlib-ng/zlib-ng/pull/1774 ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [ ] There is a reference to the original bug report and related work - [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [ ] The feature is well documented and sample code can be built with the project CMake
65 lines
1.9 KiB
C
65 lines
1.9 KiB
C
/* compare256_avx2.c -- AVX2 version of compare256
|
|
* Copyright Mika T. Lindqvist <postmaster@raasu.org>
|
|
* For conditions of distribution and use, see copyright notice in zlib.h
|
|
*/
|
|
|
|
#include "zbuild.h"
|
|
#include "zutil_p.h"
|
|
#include "deflate.h"
|
|
#include "fallback_builtins.h"
|
|
|
|
#if defined(X86_AVX2) && defined(HAVE_BUILTIN_CTZ)
|
|
|
|
#include <immintrin.h>
|
|
#ifdef _MSC_VER
|
|
# include <nmmintrin.h>
|
|
#endif
|
|
|
|
static inline uint32_t compare256_avx2_static(const uint8_t *src0, const uint8_t *src1) {
|
|
uint32_t len = 0;
|
|
|
|
do {
|
|
__m256i ymm_src0, ymm_src1, ymm_cmp;
|
|
ymm_src0 = _mm256_loadu_si256((__m256i*)src0);
|
|
ymm_src1 = _mm256_loadu_si256((__m256i*)src1);
|
|
ymm_cmp = _mm256_cmpeq_epi8(ymm_src0, ymm_src1); /* non-identical bytes = 00, identical bytes = FF */
|
|
unsigned mask = (unsigned)_mm256_movemask_epi8(ymm_cmp);
|
|
if (mask != 0xFFFFFFFF) {
|
|
uint32_t match_byte = (uint32_t)__builtin_ctz(~mask); /* Invert bits so identical = 0 */
|
|
return len + match_byte;
|
|
}
|
|
|
|
src0 += 32, src1 += 32, len += 32;
|
|
|
|
ymm_src0 = _mm256_loadu_si256((__m256i*)src0);
|
|
ymm_src1 = _mm256_loadu_si256((__m256i*)src1);
|
|
ymm_cmp = _mm256_cmpeq_epi8(ymm_src0, ymm_src1);
|
|
mask = (unsigned)_mm256_movemask_epi8(ymm_cmp);
|
|
if (mask != 0xFFFFFFFF) {
|
|
uint32_t match_byte = (uint32_t)__builtin_ctz(~mask);
|
|
return len + match_byte;
|
|
}
|
|
|
|
src0 += 32, src1 += 32, len += 32;
|
|
} while (len < 256);
|
|
|
|
return 256;
|
|
}
|
|
|
|
Z_INTERNAL uint32_t compare256_avx2(const uint8_t *src0, const uint8_t *src1) {
|
|
return compare256_avx2_static(src0, src1);
|
|
}
|
|
|
|
#define LONGEST_MATCH longest_match_avx2
|
|
#define COMPARE256 compare256_avx2_static
|
|
|
|
#include "match_tpl.h"
|
|
|
|
#define LONGEST_MATCH_SLOW
|
|
#define LONGEST_MATCH longest_match_slow_avx2
|
|
#define COMPARE256 compare256_avx2_static
|
|
|
|
#include "match_tpl.h"
|
|
|
|
#endif
|