github-mirrors/FFmpeg
1
0
Fork 0
mirror of https://mirror.skon.top/https://github.com/FFmpeg/FFmpeg synced 2026-04-20 12:50:49 +08:00
Code Issues Packages Projects Releases Wiki Activity

avutil/arm: Add armv6 optimizations for pixelutils

Browse Source 
Adapted from the corresponding me_cmp code.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This commit is contained in:
Andreas Rheinhardt
2026-03-04 23:33:22 +01:00
parent 9b84b8682f
commit 471db1d323
4 changed files with 132 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
libavutil/arm/Makefile
View File

@@ -1,6 +1,8 @@
OBJS += arm/cpu.o \
arm/float_dsp_init_arm.o \
ARMV6-OBJS-$(CONFIG_PIXELUTILS) += arm/pixelutils_armv6.o
VFP-OBJS += arm/float_dsp_init_vfp.o \
arm/float_dsp_vfp.o \

46
libavutil/arm/pixelutils.h Normal file
View File

@@ -0,0 +1,46 @@
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef AVUTIL_ARM_PIXELUTILS_H
#define AVUTIL_ARM_PIXELUTILS_H
#include <stddef.h>
#include <stdint.h>
#include "cpu.h"
#include "libavutil/attributes.h"
#include "libavutil/cpu.h"
#include "libavutil/pixelutils.h"
int ff_pixelutils_sad16_armv6(const uint8_t *src1, ptrdiff_t stride1,
const uint8_t *src2, ptrdiff_t stride2);
int ff_pixelutils_sad8_armv6 (const uint8_t *src1, ptrdiff_t stride1,
const uint8_t *src2, ptrdiff_t stride2);
static inline av_cold void ff_pixelutils_sad_init_arm(av_pixelutils_sad_fn *sad, int aligned)
{
int cpu_flags = av_get_cpu_flags();
if (have_armv6(cpu_flags)) {
if (aligned != 0) {
sad[2] = ff_pixelutils_sad8_armv6;
sad[3] = ff_pixelutils_sad16_armv6;
}
}
}
#endif

80
libavutil/arm/pixelutils_armv6.S Normal file
View File

@@ -0,0 +1,80 @@
/*
* Copyright (c) 2009 Mans Rullgard <mans@mansr.com>
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "asm.S"
function ff_pixelutils_sad16_armv6, export=1
push {r4-r10, lr}
mov r12, #0
mov r10, #16
mov lr, #0
ldm r0, {r4-r7}
ldr r8, [r2]
1:
ldr r9, [r2, #4]
pld [r0, r1]
usada8 r12, r4, r8, r12
ldr r8, [r2, #8]
pld [r2, r3]
usada8 lr, r5, r9, lr
ldr r9, [r2, #12]
usada8 r12, r6, r8, r12
subs r10, r10, #1
usada8 lr, r7, r9, lr
beq 2f
add r0, r0, r1
ldm r0, {r4-r7}
add r2, r2, r3
ldr r8, [r2]
b 1b
2:
add r0, r12, lr
pop {r4-r10, pc}
endfunc
function ff_pixelutils_sad8_armv6, export=1
pld [r2, r3]
push {r4-r10, lr}
mov r10, #8
mov r12, #0
mov lr, #0
ldrd_post r4, r5, r0, r1
1:
subs r10, r10, #2
ldr r7, [r2, #4]
ldr_post r6, r2, r3
ldrd_post r8, r9, r0, r1
usada8 r12, r4, r6, r12
pld [r2, r3]
usada8 lr, r5, r7, lr
ldr r7, [r2, #4]
ldr_post r6, r2, r3
beq 2f
ldrd_post r4, r5, r0, r1
usada8 r12, r8, r6, r12
pld [r2, r3]
usada8 lr, r9, r7, lr
b 1b
2:
usada8 r12, r8, r6, r12
usada8 lr, r9, r7, lr
add r0, r12, lr
pop {r4-r10, pc}
endfunc

4
libavutil/pixelutils.c
View File

@@ -30,6 +30,8 @@
#if ARCH_AARCH64 && HAVE_NEON
#include "aarch64/pixelutils.h"
#elif ARCH_ARM && HAVE_ARMV6
#include "arm/pixelutils.h"
#elif ARCH_RISCV
#include "riscv/pixelutils.h"
#elif ARCH_X86 && HAVE_X86ASM
@@ -94,6 +96,8 @@ av_pixelutils_sad_fn av_pixelutils_get_sad_fn(int w_bits, int h_bits, int aligne
#if ARCH_AARCH64 && HAVE_NEON
ff_pixelutils_sad_init_aarch64(sad, aligned);
#elif ARCH_ARM
ff_pixelutils_sad_init_arm(sad, aligned);
#elif ARCH_RISCV
ff_pixelutils_init_riscv(sad, aligned);
#elif ARCH_X86 && HAVE_X86ASM