coffee/libyuv
1
0
Fork 0
mirror of https://chromium.googlesource.com/libyuv/libyuv synced 2025-12-07 17:26:49 +08:00
Code Issues Packages Projects Releases Wiki Activity

ARGBAdd ported AVX2 ported to GCC.

Browse Source 
BUG=269
TESTED=try bots
R=harryjin@google.com

Review URL: https://webrtc-codereview.appspot.com/32449004

git-svn-id: http://libyuv.googlecode.com/svn/trunk@1158 16f28f9a-4ce2-e073-06de-1de4eb20be90
This commit is contained in:
fbarchard@google.com 2014-11-11 19:01:29 +00:00
parent 10d9c0d0a7
commit 9e9e26d60a
4 changed files with 35 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
README.chromium
View File

@ -1,6 +1,6 @@
Name: libyuv
URL: http://code.google.com/p/libyuv/
Version: 1157
Version: 1158
License: BSD
License File: LICENSE

5
include/libyuv/row.h
View File

@ -185,7 +185,6 @@ extern "C" {
// The code supports NaCL but requires a new compiler and validator.
#if !defined(LIBYUV_DISABLE_X86) && (defined(VISUALC_HAS_AVX2) || \
defined(CLANG_HAS_AVX2) || defined(GCC_HAS_AVX2))
// Effects:
#define HAS_COPYROW_AVX
#define HAS_ARGBPOLYNOMIALROW_AVX2
#define HAS_ARGBSHUFFLEROW_AVX2
@ -200,6 +199,9 @@ extern "C" {
#define HAS_UYVYTOUVROW_AVX2
#define HAS_SPLITUVROW_AVX2
#define HAS_MERGEUVROW_AVX2
// Effects:
#define HAS_ARGBADDROW_AVX2
#endif
// The following are require VS2012.
@ -215,7 +217,6 @@ extern "C" {
#define HAS_MIRRORROW_AVX2
// Effects:
#define HAS_ARGBADDROW_AVX2
#define HAS_ARGBATTENUATEROW_AVX2
#define HAS_ARGBMIRRORROW_AVX2
#define HAS_ARGBMULTIPLYROW_AVX2

2
include/libyuv/version.h
View File

@ -11,6 +11,6 @@
#ifndef INCLUDE_LIBYUV_VERSION_H_ // NOLINT
#define INCLUDE_LIBYUV_VERSION_H_
#define LIBYUV_VERSION 1157
#define LIBYUV_VERSION 1158
#endif // INCLUDE_LIBYUV_VERSION_H_ NOLINT

30
source/row_posix.cc
View File

@ -3955,6 +3955,36 @@ void ARGBAddRow_SSE2(const uint8* src_argb0, const uint8* src_argb1,
}
#endif // HAS_ARGBADDROW_SSE2
#ifdef HAS_ARGBADDROW_AVX2
// Add 2 rows of ARGB pixels together, 4 pixels at a time.
void ARGBAddRow_AVX2(const uint8* src_argb0, const uint8* src_argb1,
uint8* dst_argb, int width) {
asm volatile (
// 4 pixel loop.
LABELALIGN
"1: \n"
"vmovdqu " MEMACCESS(0) ",%%ymm0 \n"
"lea " MEMLEA(0x20,0) ",%0 \n"
"vpaddusb " MEMACCESS(1) ",%%ymm0,%%ymm0 \n"
"lea " MEMLEA(0x20,1) ",%1 \n"
"vmovdqu %%ymm0," MEMACCESS(2) " \n"
"lea " MEMLEA(0x20,2) ",%2 \n"
"sub $0x8,%3 \n"
"jg 1b \n"
"vzeroupper \n"
: "+r"(src_argb0), // %0
"+r"(src_argb1), // %1
"+r"(dst_argb), // %2
"+r"(width) // %3
:
: "memory", "cc"
#if defined(__SSE2__)
, "xmm0"
#endif
);
}
#endif // HAS_ARGBADDROW_AVX2
#ifdef HAS_ARGBSUBTRACTROW_SSE2
// Subtract 2 rows of ARGB pixels, 4 pixels at a time.
void ARGBSubtractRow_SSE2(const uint8* src_argb0, const uint8* src_argb1,