generic-library/ffmpeg
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

avcodec: loongson optimize blockdsp with mmi

Browse Source 
Signed-off-by: ZhouXiaoyong <zhouxiaoyong@loongson.cn>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
This commit is contained in:
周晓勇
2015-07-21 21:29:11 +08:00
committed by Michael Niedermayer
parent 39867f3e09
commit c303ad85a9
4 changed files with 170 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
libavcodec/mips/Makefile
View File

@@ -67,3 +67,4 @@ MMI-OBJS-$(CONFIG_MPEGVIDEO) += mips/mpegvideo_mmi.o
MMI-OBJS-$(CONFIG_IDCTDSP) += mips/idctdsp_mmi.o \ MMI-OBJS-$(CONFIG_IDCTDSP) += mips/idctdsp_mmi.o \
mips/simple_idct_mmi.o mips/simple_idct_mmi.o
MMI-OBJS-$(CONFIG_MPEG4_DECODER) += mips/xvid_idct_mmi.o MMI-OBJS-$(CONFIG_MPEG4_DECODER) += mips/xvid_idct_mmi.o
MMI-OBJS-$(CONFIG_BLOCKDSP) += mips/blockdsp_mmi.o

16
libavcodec/mips/blockdsp_init_mips.c
View File

@@ -1,5 +1,6 @@
/* /*
* Copyright (c) 2015 Parag Salasakar (parag.salasakar@imgtec.com) * Copyright (c) 2015 Parag Salasakar (parag.salasakar@imgtec.com)
* Zhou Xiaoyong <zhouxiaoyong@loongson.cn>
* *
* This file is part of FFmpeg. * This file is part of FFmpeg.
* *
@@ -32,9 +33,24 @@ static av_cold void blockdsp_init_msa(BlockDSPContext *c,
} }
#endif // #if HAVE_MSA #endif // #if HAVE_MSA
#if HAVE_MMI
static av_cold void blockdsp_init_mmi(BlockDSPContext *c,
unsigned high_bit_depth)
{
c->clear_block = ff_clear_block_mmi;
c->clear_blocks = ff_clear_blocks_mmi;
c->fill_block_tab[0] = ff_fill_block16_mmi;
c->fill_block_tab[1] = ff_fill_block8_mmi;
}
#endif /* HAVE_MMI */
void ff_blockdsp_init_mips(BlockDSPContext *c, unsigned high_bit_depth) void ff_blockdsp_init_mips(BlockDSPContext *c, unsigned high_bit_depth)
{ {
#if HAVE_MSA #if HAVE_MSA
blockdsp_init_msa(c, high_bit_depth); blockdsp_init_msa(c, high_bit_depth);
#endif // #if HAVE_MSA #endif // #if HAVE_MSA
#if HAVE_MMI
blockdsp_init_mmi(c, high_bit_depth);
#endif /* HAVE_MMI */
} }

6
libavcodec/mips/blockdsp_mips.h
View File

@@ -1,5 +1,6 @@
/* /*
* Copyright (c) 2015 Parag Salasakar (parag.salasakar@imgtec.com) * Copyright (c) 2015 Parag Salasakar (parag.salasakar@imgtec.com)
* Zhou Xiaoyong <zhouxiaoyong@loongson.cn>
* *
* This file is part of FFmpeg. * This file is part of FFmpeg.
* *
@@ -28,4 +29,9 @@ void ff_fill_block8_msa(uint8_t *src, uint8_t val, int stride, int height);
void ff_clear_block_msa(int16_t *block); void ff_clear_block_msa(int16_t *block);
void ff_clear_blocks_msa(int16_t *block); void ff_clear_blocks_msa(int16_t *block);
void ff_fill_block16_mmi(uint8_t *block, uint8_t value, int line_size, int h);
void ff_fill_block8_mmi(uint8_t *block, uint8_t value, int line_size, int h);
void ff_clear_block_mmi(int16_t *block);
void ff_clear_blocks_mmi(int16_t *block);
#endif // #ifndef AVCODEC_MIPS_BLOCKDSP_MIPS_H #endif // #ifndef AVCODEC_MIPS_BLOCKDSP_MIPS_H

147
libavcodec/mips/blockdsp_mmi.c Normal file
View File

@@ -0,0 +1,147 @@
/*
* Loongson SIMD optimized blockdsp
*
* Copyright (c) 2015 Loongson Technology Corporation Limited
* Copyright (c) 2015 Zhou Xiaoyong <zhouxiaoyong@loongson.cn>
*
* 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 "blockdsp_mips.h"
void ff_fill_block16_mmi(uint8_t *block, uint8_t value, int line_size, int h)
{
__asm__ volatile (
"move $8, %3 \r\n"
"move $9, %0 \r\n"
"dmtc1 %1, $f2 \r\n"
"punpcklbh $f2, $f2, $f2 \r\n"
"punpcklbh $f2, $f2, $f2 \r\n"
"punpcklbh $f2, $f2, $f2 \r\n"
"1: \r\n"
"gssdlc1 $f2, 7($9) \r\n"
"gssdrc1 $f2, 0($9) \r\n"
"gssdlc1 $f2, 15($9) \r\n"
"gssdrc1 $f2, 8($9) \r\n"
"daddi $8, $8, -1 \r\n"
"daddu $9, $9, %2 \r\n"
"bnez $8, 1b \r\n"
::"r"(block),"r"(value),"r"(line_size),"r"(h)
: "$8","$9"
);
}
void ff_fill_block8_mmi(uint8_t *block, uint8_t value, int line_size, int h)
{
__asm__ volatile (
"move $8, %3 \r\n"
"move $9, %0 \r\n"
"dmtc1 %1, $f2 \r\n"
"punpcklbh $f2, $f2, $f2 \r\n"
"punpcklbh $f2, $f2, $f2 \r\n"
"punpcklbh $f2, $f2, $f2 \r\n"
"1: \r\n"
"gssdlc1 $f2, 7($9) \r\n"
"gssdrc1 $f2, 0($9) \r\n"
"daddi $8, $8, -1 \r\n"
"daddu $9, $9, %2 \r\n"
"bnez $8, 1b \r\n"
::"r"(block),"r"(value),"r"(line_size),"r"(h)
: "$8","$9"
);
}
void ff_clear_block_mmi(int16_t *block)
{
__asm__ volatile (
"xor $f0, $f0, $f0 \r\n"
"xor $f2, $f2, $f2 \r\n"
"gssqc1 $f0, $f2, 0(%0) \r\n"
"gssqc1 $f0, $f2, 16(%0) \r\n"
"gssqc1 $f0, $f2, 32(%0) \r\n"
"gssqc1 $f0, $f2, 48(%0) \r\n"
"gssqc1 $f0, $f2, 64(%0) \r\n"
"gssqc1 $f0, $f2, 80(%0) \r\n"
"gssqc1 $f0, $f2, 96(%0) \r\n"
"gssqc1 $f0, $f2, 112(%0) \r\n"
::"r"(block)
: "memory"
);
}
void ff_clear_blocks_mmi(int16_t *block)
{
__asm__ volatile (
"xor $f0, $f0, $f0 \r\n"
"xor $f2, $f2, $f2 \r\n"
"gssqc1 $f0, $f2, 0(%0) \r\n"
"gssqc1 $f0, $f2, 16(%0) \r\n"
"gssqc1 $f0, $f2, 32(%0) \r\n"
"gssqc1 $f0, $f2, 48(%0) \r\n"
"gssqc1 $f0, $f2, 64(%0) \r\n"
"gssqc1 $f0, $f2, 80(%0) \r\n"
"gssqc1 $f0, $f2, 96(%0) \r\n"
"gssqc1 $f0, $f2, 112(%0) \r\n"
"gssqc1 $f0, $f2, 128(%0) \r\n"
"gssqc1 $f0, $f2, 144(%0) \r\n"
"gssqc1 $f0, $f2, 160(%0) \r\n"
"gssqc1 $f0, $f2, 176(%0) \r\n"
"gssqc1 $f0, $f2, 192(%0) \r\n"
"gssqc1 $f0, $f2, 208(%0) \r\n"
"gssqc1 $f0, $f2, 224(%0) \r\n"
"gssqc1 $f0, $f2, 240(%0) \r\n"
"gssqc1 $f0, $f2, 256(%0) \r\n"
"gssqc1 $f0, $f2, 272(%0) \r\n"
"gssqc1 $f0, $f2, 288(%0) \r\n"
"gssqc1 $f0, $f2, 304(%0) \r\n"
"gssqc1 $f0, $f2, 320(%0) \r\n"
"gssqc1 $f0, $f2, 336(%0) \r\n"
"gssqc1 $f0, $f2, 352(%0) \r\n"
"gssqc1 $f0, $f2, 368(%0) \r\n"
"gssqc1 $f0, $f2, 384(%0) \r\n"
"gssqc1 $f0, $f2, 400(%0) \r\n"
"gssqc1 $f0, $f2, 416(%0) \r\n"
"gssqc1 $f0, $f2, 432(%0) \r\n"
"gssqc1 $f0, $f2, 448(%0) \r\n"
"gssqc1 $f0, $f2, 464(%0) \r\n"
"gssqc1 $f0, $f2, 480(%0) \r\n"
"gssqc1 $f0, $f2, 496(%0) \r\n"
"gssqc1 $f0, $f2, 512(%0) \r\n"
"gssqc1 $f0, $f2, 528(%0) \r\n"
"gssqc1 $f0, $f2, 544(%0) \r\n"
"gssqc1 $f0, $f2, 560(%0) \r\n"
"gssqc1 $f0, $f2, 576(%0) \r\n"
"gssqc1 $f0, $f2, 592(%0) \r\n"
"gssqc1 $f0, $f2, 608(%0) \r\n"
"gssqc1 $f0, $f2, 624(%0) \r\n"
"gssqc1 $f0, $f2, 640(%0) \r\n"
"gssqc1 $f0, $f2, 656(%0) \r\n"
"gssqc1 $f0, $f2, 672(%0) \r\n"
"gssqc1 $f0, $f2, 688(%0) \r\n"
"gssqc1 $f0, $f2, 704(%0) \r\n"
"gssqc1 $f0, $f2, 720(%0) \r\n"
"gssqc1 $f0, $f2, 736(%0) \r\n"
"gssqc1 $f0, $f2, 752(%0) \r\n"
::"r"(block)
: "memory"
);
}