RepoMirrors
/
ffmpeg
mirror of https://git.ffmpeg.org/ffmpeg.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for mpegvideoencdsp functions 

This patch adds MSA (MIPS-SIMD-Arch) optimizations for mpegvideoencdsp functions in new file mpegvideoencdsp_msa.c

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
Shivraj Patil 2015-06-29 20:57:13 +05:30 committed by Michael Niedermayer
parent 115c14c3b6
commit 2f3f98af2b
6 changed files with 142 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
libavcodec/mips/Makefile
View File

@ -30,6 +30,7 @@ OBJS-$(CONFIG_HPELDSP) += mips/hpeldsp_init_mips.o
OBJS-$(CONFIG_BLOCKDSP) += mips/blockdsp_init_mips.o
OBJS-$(CONFIG_PIXBLOCKDSP) += mips/pixblockdsp_init_mips.o
OBJS-$(CONFIG_MPEGVIDEO) += mips/mpegvideo_init_mips.o
OBJS-$(CONFIG_MPEGVIDEOENC) += mips/mpegvideoencdsp_init_mips.o
MSA-OBJS-$(CONFIG_HEVC_DECODER) += mips/hevcdsp_msa.o \
mips/hevc_mc_uni_msa.o \
mips/hevc_mc_uniw_msa.o \
@ -49,5 +50,6 @@ MSA-OBJS-$(CONFIG_HPELDSP) += mips/hpeldsp_msa.o
MSA-OBJS-$(CONFIG_BLOCKDSP) += mips/blockdsp_msa.o
MSA-OBJS-$(CONFIG_PIXBLOCKDSP) += mips/pixblockdsp_msa.o
MSA-OBJS-$(CONFIG_MPEGVIDEO) += mips/mpegvideo_msa.o
MSA-OBJS-$(CONFIG_MPEGVIDEOENC) += mips/mpegvideoencdsp_msa.o
LOONGSON3-OBJS-$(CONFIG_H264DSP) += mips/h264dsp_mmi.o
LOONGSON3-OBJS-$(CONFIG_H264CHROMA) += mips/h264chroma_mmi.o

40
libavcodec/mips/mpegvideoencdsp_init_mips.c Normal file
View File

@ -0,0 +1,40 @@
/*
* Copyright (c) 2015 Manojkumar Bhosale (Manojkumar.Bhosale@imgtec.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 "libavcodec/bit_depth_template.c"
#include "h263dsp_mips.h"
#if HAVE_MSA
static av_cold void mpegvideoencdsp_init_msa(MpegvideoEncDSPContext *c,
AVCodecContext *avctx)
{
#if BIT_DEPTH == 8
c->pix_sum = ff_pix_sum_msa;
#endif
}
#endif // #if HAVE_MSA
av_cold void ff_mpegvideoencdsp_init_mips(MpegvideoEncDSPContext *c,
AVCodecContext *avctx)
{
#if HAVE_MSA
mpegvideoencdsp_init_msa(c, avctx);
#endif // #if HAVE_MSA
}

62
libavcodec/mips/mpegvideoencdsp_msa.c Normal file
View File

@ -0,0 +1,62 @@
/*
* Copyright (c) 2015 Manojkumar Bhosale (Manojkumar.Bhosale@imgtec.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 "h263dsp_mips.h"
#include "libavutil/mips/generic_macros_msa.h"
static int32_t sum_u8src_16width_msa(uint8_t *src, int32_t stride)
{
uint32_t sum = 0;
v16u8 in0, in1, in2, in3, in4, in5, in6, in7;
v16u8 in8, in9, in10, in11, in12, in13, in14, in15;
LD_UB8(src, stride, in0, in1, in2, in3, in4, in5, in6, in7);
src += (8 * stride);
LD_UB8(src, stride, in8, in9, in10, in11, in12, in13, in14, in15);
HADD_UB4_UB(in0, in1, in2, in3, in0, in1, in2, in3);
HADD_UB4_UB(in4, in5, in6, in7, in4, in5, in6, in7);
HADD_UB4_UB(in8, in9, in10, in11, in8, in9, in10, in11);
HADD_UB4_UB(in12, in13, in14, in15, in12, in13, in14, in15);
sum = HADD_UH_U32(in0);
sum += HADD_UH_U32(in1);
sum += HADD_UH_U32(in2);
sum += HADD_UH_U32(in3);
sum += HADD_UH_U32(in4);
sum += HADD_UH_U32(in5);
sum += HADD_UH_U32(in6);
sum += HADD_UH_U32(in7);
sum += HADD_UH_U32(in8);
sum += HADD_UH_U32(in9);
sum += HADD_UH_U32(in10);
sum += HADD_UH_U32(in11);
sum += HADD_UH_U32(in12);
sum += HADD_UH_U32(in13);
sum += HADD_UH_U32(in14);
sum += HADD_UH_U32(in15);
return sum;
}
int ff_pix_sum_msa(uint8_t *pix, int line_size)
{
return sum_u8src_16width_msa(pix, line_size);
}

2
libavcodec/mpegvideoencdsp.c
View File

@ -176,4 +176,6 @@ av_cold void ff_mpegvideoencdsp_init(MpegvideoEncDSPContext *c,
ff_mpegvideoencdsp_init_ppc(c, avctx);
if (ARCH_X86)
ff_mpegvideoencdsp_init_x86(c, avctx);
if (ARCH_MIPS)
ff_mpegvideoencdsp_init_mips(c, avctx);
}

2
libavcodec/mpegvideoencdsp.h
View File

@ -52,5 +52,7 @@ void ff_mpegvideoencdsp_init_ppc(MpegvideoEncDSPContext *c,
AVCodecContext *avctx);
void ff_mpegvideoencdsp_init_x86(MpegvideoEncDSPContext *c,
AVCodecContext *avctx);
void ff_mpegvideoencdsp_init_mips(MpegvideoEncDSPContext *c,
AVCodecContext *avctx);
#endif /* AVCODEC_MPEGVIDEOENCDSP_H */

34
libavutil/mips/generic_macros_msa.h
View File

@ -1262,6 +1262,15 @@
}
#define HADD_UB3_UH(...) HADD_UB3(v8u16, __VA_ARGS__)
#define HADD_UB4(RTYPE, in0, in1, in2, in3, out0, out1, out2, out3) \
{ \
HADD_UB2(RTYPE, in0, in1, out0, out1); \
HADD_UB2(RTYPE, in2, in3, out2, out3); \
}
#define HADD_UB4_UB(...) HADD_UB4(v16u8, __VA_ARGS__)
#define HADD_UB4_UH(...) HADD_UB4(v8u16, __VA_ARGS__)
#define HADD_UB4_SH(...) HADD_UB4(v8i16, __VA_ARGS__)
/* Description : Horizontal subtraction of unsigned byte vector elements
Arguments : Inputs - in0, in1
Outputs - out0, out1
@ -1771,6 +1780,15 @@
#define SPLATI_H2_SB(...) SPLATI_H2(v16i8, __VA_ARGS__)
#define SPLATI_H2_SH(...) SPLATI_H2(v8i16, __VA_ARGS__)
#define SPLATI_H3(RTYPE, in, idx0, idx1, idx2, \
out0, out1, out2) \
{ \
SPLATI_H2(RTYPE, in, idx0, idx1, out0, out1); \
out2 = (RTYPE) __msa_splati_h((v8i16) in, idx2); \
}
#define SPLATI_H3_SB(...) SPLATI_H3(v16i8, __VA_ARGS__)
#define SPLATI_H3_SH(...) SPLATI_H3(v8i16, __VA_ARGS__)
#define SPLATI_H4(RTYPE, in, idx0, idx1, idx2, idx3, \
out0, out1, out2, out3) \
{ \
@ -2823,4 +2841,20 @@
tmp_m = __msa_pckev_b((v16i8) in1, (v16i8) in0); \
ST_SB(tmp_m, (pdst)); \
}
/* Description : Horizontal 2 tap filter kernel code
Arguments : Inputs - in0, in1, mask, coeff, shift
*/
#define HORIZ_2TAP_FILT_UH(in0, in1, mask, coeff, shift) \
( { \
v16i8 tmp0_m; \
v8u16 tmp1_m; \
\
tmp0_m = __msa_vshf_b((v16i8) mask, (v16i8) in1, (v16i8) in0); \
tmp1_m = __msa_dotp_u_h((v16u8) tmp0_m, (v16u8) coeff); \
tmp1_m = (v8u16) __msa_srari_h((v8i16) tmp1_m, shift); \
tmp1_m = __msa_sat_u_h(tmp1_m, shift); \
\
tmp1_m; \
} )
#endif /* AVUTIL_MIPS_GENERIC_MACROS_MSA_H */