Commit 012ba786 authored by Shivraj Patil's avatar Shivraj Patil Committed by Michael Niedermayer

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for VP9 MC functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for VP9 MC functions in new file vp9_mc_msa.c
Signed-off-by: 's avatarShivraj Patil <shivraj.patil@imgtec.com>
Reviewed-by: 's avatar"Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: 's avatarMichael Niedermayer <michael@niedermayer.cc>
parent 2af180bf
......@@ -20,6 +20,7 @@ MIPSDSPR1-OBJS-$(CONFIG_AAC_ENCODER) += mips/aaccoder_mips.o
MIPSFPU-OBJS-$(CONFIG_AAC_ENCODER) += mips/iirfilter_mips.o
OBJS-$(CONFIG_HEVC_DECODER) += mips/hevcdsp_init_mips.o \
mips/hevcpred_init_mips.o
OBJS-$(CONFIG_VP9_DECODER) += mips/vp9dsp_init_mips.o
OBJS-$(CONFIG_H264DSP) += mips/h264dsp_init_mips.o
OBJS-$(CONFIG_H264QPEL) += mips/h264qpel_init_mips.o
OBJS-$(CONFIG_H264CHROMA) += mips/h264chroma_init_mips.o
......@@ -41,6 +42,7 @@ MSA-OBJS-$(CONFIG_HEVC_DECODER) += mips/hevcdsp_msa.o \
mips/hevc_idct_msa.o \
mips/hevc_lpf_sao_msa.o \
mips/hevcpred_msa.o
MSA-OBJS-$(CONFIG_VP9_DECODER) += mips/vp9_mc_msa.o
MSA-OBJS-$(CONFIG_H264DSP) += mips/h264dsp_msa.o \
mips/h264idct_msa.o
MSA-OBJS-$(CONFIG_H264QPEL) += mips/h264qpel_msa.o
......
This source diff could not be displayed because it is too large. You can view the blob instead.
/*
* Copyright (c) 2015 Shivraj Patil (Shivraj.Patil@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 "config.h"
#include "libavutil/common.h"
#include "libavcodec/vp9dsp.h"
#include "vp9dsp_mips.h"
#if HAVE_MSA
static av_cold void vp9dsp_mc_init_msa(VP9DSPContext *dsp, int bpp)
{
if (bpp == 8) {
#define init_fpel(idx1, idx2, sz, type) \
dsp->mc[idx1][FILTER_8TAP_SMOOTH ][idx2][0][0] = ff_##type##sz##_msa; \
dsp->mc[idx1][FILTER_8TAP_REGULAR][idx2][0][0] = ff_##type##sz##_msa; \
dsp->mc[idx1][FILTER_8TAP_SHARP ][idx2][0][0] = ff_##type##sz##_msa; \
dsp->mc[idx1][FILTER_BILINEAR ][idx2][0][0] = ff_##type##sz##_msa
#define init_copy_avg(idx, sz) \
init_fpel(idx, 0, sz, copy); \
init_fpel(idx, 1, sz, avg)
#define init_avg(idx, sz) \
init_fpel(idx, 1, sz, avg)
init_copy_avg(0, 64);
init_copy_avg(1, 32);
init_copy_avg(2, 16);
init_copy_avg(3, 8);
init_avg(4, 4);
#undef init_copy_avg
#undef init_avg
#undef init_fpel
#define init_subpel1(idx1, idx2, idxh, idxv, sz, dir, type) \
dsp->mc[idx1][FILTER_8TAP_SMOOTH ][idx2][idxh][idxv] = \
ff_##type##_8tap_smooth_##sz##dir##_msa; \
dsp->mc[idx1][FILTER_8TAP_REGULAR][idx2][idxh][idxv] = \
ff_##type##_8tap_regular_##sz##dir##_msa; \
dsp->mc[idx1][FILTER_8TAP_SHARP ][idx2][idxh][idxv] = \
ff_##type##_8tap_sharp_##sz##dir##_msa;
#define init_subpel2(idx, idxh, idxv, dir, type) \
init_subpel1(0, idx, idxh, idxv, 64, dir, type); \
init_subpel1(1, idx, idxh, idxv, 32, dir, type); \
init_subpel1(2, idx, idxh, idxv, 16, dir, type); \
init_subpel1(3, idx, idxh, idxv, 8, dir, type); \
init_subpel1(4, idx, idxh, idxv, 4, dir, type)
#define init_subpel3(idx, type) \
init_subpel2(idx, 1, 1, hv, type); \
init_subpel2(idx, 0, 1, v, type); \
init_subpel2(idx, 1, 0, h, type)
init_subpel3(0, put);
init_subpel3(1, avg);
#undef init_subpel1
#undef init_subpel2
#undef init_subpel3
}
}
static av_cold void vp9dsp_init_msa(VP9DSPContext *dsp, int bpp)
{
vp9dsp_mc_init_msa(dsp, bpp);
}
#endif // #if HAVE_MSA
av_cold void ff_vp9dsp_init_mips(VP9DSPContext *dsp, int bpp)
{
#if HAVE_MSA
vp9dsp_init_msa(dsp, bpp);
#endif // #if HAVE_MSA
}
/*
* Copyright (c) 2015 Shivraj Patil (Shivraj.Patil@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
*/
#ifndef AVCODEC_MIPS_VP9DSP_MIPS_H
#define AVCODEC_MIPS_VP9DSP_MIPS_H
#define VP9_8TAP_MIPS_MSA_FUNC(SIZE, type, type_idx) \
void ff_put_8tap_##type##_##SIZE##h_msa(uint8_t *dst, ptrdiff_t dststride, \
const uint8_t *src, \
ptrdiff_t srcstride, \
int h, int mx, int my); \
\
void ff_put_8tap_##type##_##SIZE##v_msa(uint8_t *dst, ptrdiff_t dststride, \
const uint8_t *src, \
ptrdiff_t srcstride, \
int h, int mx, int my); \
\
void ff_put_8tap_##type##_##SIZE##hv_msa(uint8_t *dst, ptrdiff_t dststride, \
const uint8_t *src, \
ptrdiff_t srcstride, \
int h, int mx, int my); \
\
void ff_avg_8tap_##type##_##SIZE##h_msa(uint8_t *dst, ptrdiff_t dststride, \
const uint8_t *src, \
ptrdiff_t srcstride, \
int h, int mx, int my); \
\
void ff_avg_8tap_##type##_##SIZE##v_msa(uint8_t *dst, ptrdiff_t dststride, \
const uint8_t *src, \
ptrdiff_t srcstride, \
int h, int mx, int my); \
\
void ff_avg_8tap_##type##_##SIZE##hv_msa(uint8_t *dst, ptrdiff_t dststride, \
const uint8_t *src, \
ptrdiff_t srcstride, \
int h, int mx, int my);
#define VP9_COPY_AVG_MIPS_MSA_FUNC(SIZE) \
void ff_copy##SIZE##_msa(uint8_t *dst, ptrdiff_t dststride, \
const uint8_t *src, ptrdiff_t srcstride, \
int h, int mx, int my); \
\
void ff_avg##SIZE##_msa(uint8_t *dst, ptrdiff_t dststride, \
const uint8_t *src, ptrdiff_t srcstride, \
int h, int mx, int my);
VP9_8TAP_MIPS_MSA_FUNC(64, regular, FILTER_8TAP_REGULAR);
VP9_8TAP_MIPS_MSA_FUNC(32, regular, FILTER_8TAP_REGULAR);
VP9_8TAP_MIPS_MSA_FUNC(16, regular, FILTER_8TAP_REGULAR);
VP9_8TAP_MIPS_MSA_FUNC(8, regular, FILTER_8TAP_REGULAR);
VP9_8TAP_MIPS_MSA_FUNC(4, regular, FILTER_8TAP_REGULAR);
VP9_8TAP_MIPS_MSA_FUNC(64, sharp, FILTER_8TAP_SHARP);
VP9_8TAP_MIPS_MSA_FUNC(32, sharp, FILTER_8TAP_SHARP);
VP9_8TAP_MIPS_MSA_FUNC(16, sharp, FILTER_8TAP_SHARP);
VP9_8TAP_MIPS_MSA_FUNC(8, sharp, FILTER_8TAP_SHARP);
VP9_8TAP_MIPS_MSA_FUNC(4, sharp, FILTER_8TAP_SHARP);
VP9_8TAP_MIPS_MSA_FUNC(64, smooth, FILTER_8TAP_SMOOTH);
VP9_8TAP_MIPS_MSA_FUNC(32, smooth, FILTER_8TAP_SMOOTH);
VP9_8TAP_MIPS_MSA_FUNC(16, smooth, FILTER_8TAP_SMOOTH);
VP9_8TAP_MIPS_MSA_FUNC(8, smooth, FILTER_8TAP_SMOOTH);
VP9_8TAP_MIPS_MSA_FUNC(4, smooth, FILTER_8TAP_SMOOTH);
VP9_COPY_AVG_MIPS_MSA_FUNC(64);
VP9_COPY_AVG_MIPS_MSA_FUNC(32);
VP9_COPY_AVG_MIPS_MSA_FUNC(16);
VP9_COPY_AVG_MIPS_MSA_FUNC(8);
VP9_COPY_AVG_MIPS_MSA_FUNC(4);
#undef VP9_8TAP_MIPS_MSA_FUNC
#undef VP9_COPY_AVG_MIPS_MSA_FUNC
#endif // #ifndef AVCODEC_MIPS_VP9DSP_MIPS_H
......@@ -37,4 +37,5 @@ av_cold void ff_vp9dsp_init(VP9DSPContext *dsp, int bpp)
}
if (ARCH_X86) ff_vp9dsp_init_x86(dsp, bpp);
if (ARCH_MIPS) ff_vp9dsp_init_mips(dsp, bpp);
}
......@@ -127,5 +127,6 @@ void ff_vp9dsp_init_10(VP9DSPContext *dsp);
void ff_vp9dsp_init_12(VP9DSPContext *dsp);
void ff_vp9dsp_init_x86(VP9DSPContext *dsp, int bpp);
void ff_vp9dsp_init_mips(VP9DSPContext *dsp, int bpp);
#endif /* AVCODEC_VP9DSP_H */
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment