a5dfeb612e
From 52.503s (~40fps) to 27.973sec (~80fps) decoding of 480p sintel trailer, i.e. a ~2x speedup overall, on a Nexus S. Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
310 lines
14 KiB
C
310 lines
14 KiB
C
/**
|
|
* 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 <stdint.h>
|
|
#include "libavcodec/vp8dsp.h"
|
|
|
|
void ff_vp8_luma_dc_wht_dc_armv6(DCTELEM block[4][4][16], DCTELEM dc[16]);
|
|
|
|
#define idct_funcs(opt) \
|
|
void ff_vp8_luma_dc_wht_ ## opt(DCTELEM block[4][4][16], DCTELEM dc[16]); \
|
|
void ff_vp8_idct_add_ ## opt(uint8_t *dst, DCTELEM block[16], int stride); \
|
|
void ff_vp8_idct_dc_add_ ## opt(uint8_t *dst, DCTELEM block[16], int stride); \
|
|
void ff_vp8_idct_dc_add4y_ ## opt(uint8_t *dst, DCTELEM block[4][16], int stride); \
|
|
void ff_vp8_idct_dc_add4uv_ ## opt(uint8_t *dst, DCTELEM block[4][16], int stride)
|
|
|
|
idct_funcs(neon);
|
|
idct_funcs(armv6);
|
|
|
|
void ff_vp8_v_loop_filter16_neon(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I, int hev_thresh);
|
|
void ff_vp8_h_loop_filter16_neon(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I, int hev_thresh);
|
|
void ff_vp8_v_loop_filter8uv_neon(uint8_t *dstU, uint8_t *dstV, int stride,
|
|
int flim_E, int flim_I, int hev_thresh);
|
|
void ff_vp8_h_loop_filter8uv_neon(uint8_t *dstU, uint8_t *dstV, int stride,
|
|
int flim_E, int flim_I, int hev_thresh);
|
|
|
|
void ff_vp8_v_loop_filter16_inner_neon(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I, int hev_thresh);
|
|
void ff_vp8_h_loop_filter16_inner_neon(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I, int hev_thresh);
|
|
void ff_vp8_v_loop_filter8uv_inner_neon(uint8_t *dstU, uint8_t *dstV,
|
|
int stride, int flim_E, int flim_I,
|
|
int hev_thresh);
|
|
void ff_vp8_h_loop_filter8uv_inner_neon(uint8_t *dstU, uint8_t *dstV,
|
|
int stride, int flim_E, int flim_I,
|
|
int hev_thresh);
|
|
|
|
void ff_vp8_v_loop_filter_inner_armv6(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I,
|
|
int hev_thresh, int count);
|
|
void ff_vp8_h_loop_filter_inner_armv6(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I,
|
|
int hev_thresh, int count);
|
|
void ff_vp8_v_loop_filter_armv6(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I,
|
|
int hev_thresh, int count);
|
|
void ff_vp8_h_loop_filter_armv6(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I,
|
|
int hev_thresh, int count);
|
|
|
|
static void ff_vp8_v_loop_filter16_armv6(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I, int hev_thresh)
|
|
{
|
|
ff_vp8_v_loop_filter_armv6(dst, stride, flim_E, flim_I, hev_thresh, 4);
|
|
}
|
|
|
|
static void ff_vp8_h_loop_filter16_armv6(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I, int hev_thresh)
|
|
{
|
|
ff_vp8_h_loop_filter_armv6(dst, stride, flim_E, flim_I, hev_thresh, 4);
|
|
}
|
|
|
|
static void ff_vp8_v_loop_filter8uv_armv6(uint8_t *dstU, uint8_t *dstV, int stride,
|
|
int flim_E, int flim_I, int hev_thresh)
|
|
{
|
|
ff_vp8_v_loop_filter_armv6(dstU, stride, flim_E, flim_I, hev_thresh, 2);
|
|
ff_vp8_v_loop_filter_armv6(dstV, stride, flim_E, flim_I, hev_thresh, 2);
|
|
}
|
|
|
|
static void ff_vp8_h_loop_filter8uv_armv6(uint8_t *dstU, uint8_t *dstV, int stride,
|
|
int flim_E, int flim_I, int hev_thresh)
|
|
{
|
|
ff_vp8_h_loop_filter_armv6(dstU, stride, flim_E, flim_I, hev_thresh, 2);
|
|
ff_vp8_h_loop_filter_armv6(dstV, stride, flim_E, flim_I, hev_thresh, 2);
|
|
}
|
|
|
|
static void ff_vp8_v_loop_filter16_inner_armv6(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I, int hev_thresh)
|
|
{
|
|
ff_vp8_v_loop_filter_inner_armv6(dst, stride, flim_E, flim_I, hev_thresh, 4);
|
|
}
|
|
|
|
static void ff_vp8_h_loop_filter16_inner_armv6(uint8_t *dst, int stride,
|
|
int flim_E, int flim_I, int hev_thresh)
|
|
{
|
|
ff_vp8_h_loop_filter_inner_armv6(dst, stride, flim_E, flim_I, hev_thresh, 4);
|
|
}
|
|
|
|
static void ff_vp8_v_loop_filter8uv_inner_armv6(uint8_t *dstU, uint8_t *dstV,
|
|
int stride, int flim_E, int flim_I,
|
|
int hev_thresh)
|
|
{
|
|
ff_vp8_v_loop_filter_inner_armv6(dstU, stride, flim_E, flim_I, hev_thresh, 2);
|
|
ff_vp8_v_loop_filter_inner_armv6(dstV, stride, flim_E, flim_I, hev_thresh, 2);
|
|
}
|
|
|
|
static void ff_vp8_h_loop_filter8uv_inner_armv6(uint8_t *dstU, uint8_t *dstV,
|
|
int stride, int flim_E, int flim_I,
|
|
int hev_thresh)
|
|
{
|
|
ff_vp8_h_loop_filter_inner_armv6(dstU, stride, flim_E, flim_I, hev_thresh, 2);
|
|
ff_vp8_h_loop_filter_inner_armv6(dstV, stride, flim_E, flim_I, hev_thresh, 2);
|
|
}
|
|
|
|
#define simple_lf_funcs(opt) \
|
|
void ff_vp8_v_loop_filter16_simple_ ## opt(uint8_t *dst, int stride, int flim); \
|
|
void ff_vp8_h_loop_filter16_simple_ ## opt(uint8_t *dst, int stride, int flim)
|
|
|
|
simple_lf_funcs(neon);
|
|
simple_lf_funcs(armv6);
|
|
|
|
#define VP8_MC_OPT(n, opt) \
|
|
void ff_put_vp8_##n##_##opt(uint8_t *dst, int dststride, \
|
|
uint8_t *src, int srcstride, \
|
|
int h, int x, int y)
|
|
|
|
#define VP8_MC(n) \
|
|
VP8_MC_OPT(n, neon)
|
|
|
|
#define VP8_EPEL(w) \
|
|
VP8_MC(epel ## w ## _h4); \
|
|
VP8_MC(epel ## w ## _h6); \
|
|
VP8_MC(epel ## w ## _h4v4); \
|
|
VP8_MC(epel ## w ## _h6v4); \
|
|
VP8_MC(epel ## w ## _v4); \
|
|
VP8_MC(epel ## w ## _v6); \
|
|
VP8_MC(epel ## w ## _h4v6); \
|
|
VP8_MC(epel ## w ## _h6v6)
|
|
|
|
VP8_EPEL(16);
|
|
VP8_MC(pixels16);
|
|
VP8_MC_OPT(pixels16, armv6);
|
|
VP8_EPEL(8);
|
|
VP8_MC(pixels8);
|
|
VP8_MC_OPT(pixels8, armv6);
|
|
VP8_EPEL(4);
|
|
VP8_MC_OPT(pixels4, armv6);
|
|
|
|
VP8_MC(bilin16_h);
|
|
VP8_MC(bilin16_v);
|
|
VP8_MC(bilin16_hv);
|
|
VP8_MC(bilin8_h);
|
|
VP8_MC(bilin8_v);
|
|
VP8_MC(bilin8_hv);
|
|
VP8_MC(bilin4_h);
|
|
VP8_MC(bilin4_v);
|
|
VP8_MC(bilin4_hv);
|
|
|
|
#define VP8_V6_MC(n) \
|
|
void ff_put_vp8_##n##_armv6(uint8_t *dst, int dststride, uint8_t *src, \
|
|
int srcstride, int w, int h, int mxy)
|
|
|
|
VP8_V6_MC(epel_v6);
|
|
VP8_V6_MC(epel_h6);
|
|
VP8_V6_MC(epel_v4);
|
|
VP8_V6_MC(epel_h4);
|
|
VP8_V6_MC(bilin_v);
|
|
VP8_V6_MC(bilin_h);
|
|
|
|
#define VP8_EPEL_HV(SIZE, TAPNUMX, TAPNUMY, NAME, HNAME, VNAME, MAXHEIGHT) \
|
|
static void ff_put_vp8_##NAME##SIZE##_##HNAME##VNAME##_armv6( \
|
|
uint8_t *dst, int dststride, uint8_t *src, \
|
|
int srcstride, int h, int mx, int my) \
|
|
{ \
|
|
DECLARE_ALIGNED(4, uint8_t, tmp)[SIZE * (MAXHEIGHT + TAPNUMY - 1)]; \
|
|
uint8_t *tmpptr = tmp + SIZE * (TAPNUMY / 2 - 1); \
|
|
src -= srcstride * (TAPNUMY / 2 - 1); \
|
|
ff_put_vp8_ ## NAME ## _ ## HNAME ## _armv6(tmp, SIZE, src, srcstride, \
|
|
SIZE, h + TAPNUMY - 1, mx); \
|
|
ff_put_vp8_ ## NAME ## _ ## VNAME ## _armv6(dst, dststride, tmpptr, SIZE, \
|
|
SIZE, h, my); \
|
|
}
|
|
|
|
VP8_EPEL_HV(16, 6, 6, epel, h6, v6, 16);
|
|
VP8_EPEL_HV(16, 2, 2, bilin, h, v, 16);
|
|
VP8_EPEL_HV(8, 6, 6, epel, h6, v6, 16);
|
|
VP8_EPEL_HV(8, 4, 6, epel, h4, v6, 16);
|
|
VP8_EPEL_HV(8, 6, 4, epel, h6, v4, 16);
|
|
VP8_EPEL_HV(8, 4, 4, epel, h4, v4, 16);
|
|
VP8_EPEL_HV(8, 2, 2, bilin, h, v, 16);
|
|
VP8_EPEL_HV(4, 6, 6, epel, h6, v6, 8);
|
|
VP8_EPEL_HV(4, 4, 6, epel, h4, v6, 8);
|
|
VP8_EPEL_HV(4, 6, 4, epel, h6, v4, 8);
|
|
VP8_EPEL_HV(4, 4, 4, epel, h4, v4, 8);
|
|
VP8_EPEL_HV(4, 2, 2, bilin, h, v, 8);
|
|
|
|
extern void put_vp8_epel4_v6_c(uint8_t *dst, int d, uint8_t *src, int s, int h, int mx, int my);
|
|
#undef printf
|
|
#define VP8_EPEL_H_OR_V(SIZE, NAME, HV) \
|
|
static void ff_put_vp8_##NAME##SIZE##_##HV##_armv6( \
|
|
uint8_t *dst, int dststride, uint8_t *src, \
|
|
int srcstride, int h, int mx, int my) \
|
|
{ \
|
|
ff_put_vp8_## NAME ## _ ## HV ## _armv6(dst, dststride, src, srcstride, \
|
|
SIZE, h, mx | my); \
|
|
}
|
|
|
|
VP8_EPEL_H_OR_V(4, epel, h6);
|
|
VP8_EPEL_H_OR_V(4, epel, h4);
|
|
VP8_EPEL_H_OR_V(4, epel, v6);
|
|
VP8_EPEL_H_OR_V(4, epel, v4);
|
|
VP8_EPEL_H_OR_V(4, bilin, v);
|
|
VP8_EPEL_H_OR_V(4, bilin, h);
|
|
VP8_EPEL_H_OR_V(8, epel, h6);
|
|
VP8_EPEL_H_OR_V(8, epel, h4);
|
|
VP8_EPEL_H_OR_V(8, epel, v6);
|
|
VP8_EPEL_H_OR_V(8, epel, v4);
|
|
VP8_EPEL_H_OR_V(8, bilin, v);
|
|
VP8_EPEL_H_OR_V(8, bilin, h);
|
|
VP8_EPEL_H_OR_V(16, epel, h6);
|
|
VP8_EPEL_H_OR_V(16, epel, v6);
|
|
VP8_EPEL_H_OR_V(16, bilin, v);
|
|
VP8_EPEL_H_OR_V(16, bilin, h);
|
|
|
|
av_cold void ff_vp8dsp_init_arm(VP8DSPContext *dsp)
|
|
{
|
|
#define set_func_ptrs(opt) \
|
|
dsp->vp8_luma_dc_wht = ff_vp8_luma_dc_wht_##opt; \
|
|
dsp->vp8_luma_dc_wht_dc = ff_vp8_luma_dc_wht_dc_armv6; \
|
|
\
|
|
dsp->vp8_idct_add = ff_vp8_idct_add_##opt; \
|
|
dsp->vp8_idct_dc_add = ff_vp8_idct_dc_add_##opt; \
|
|
dsp->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_##opt; \
|
|
dsp->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_##opt; \
|
|
\
|
|
dsp->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16_##opt; \
|
|
dsp->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16_##opt; \
|
|
dsp->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_##opt; \
|
|
dsp->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_##opt; \
|
|
\
|
|
dsp->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16_inner_##opt; \
|
|
dsp->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16_inner_##opt; \
|
|
dsp->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_##opt; \
|
|
dsp->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_##opt; \
|
|
\
|
|
dsp->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter16_simple_##opt; \
|
|
dsp->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter16_simple_##opt; \
|
|
\
|
|
dsp->put_vp8_epel_pixels_tab[0][0][0] = ff_put_vp8_pixels16_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[0][0][2] = ff_put_vp8_epel16_h6_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[0][2][0] = ff_put_vp8_epel16_v6_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[0][2][2] = ff_put_vp8_epel16_h6v6_##opt; \
|
|
\
|
|
dsp->put_vp8_epel_pixels_tab[1][0][0] = ff_put_vp8_pixels8_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[1][0][1] = ff_put_vp8_epel8_h4_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[1][0][2] = ff_put_vp8_epel8_h6_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[1][1][0] = ff_put_vp8_epel8_v4_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[1][1][1] = ff_put_vp8_epel8_h4v4_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[1][1][2] = ff_put_vp8_epel8_h6v4_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[1][2][0] = ff_put_vp8_epel8_v6_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[1][2][1] = ff_put_vp8_epel8_h4v6_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[1][2][2] = ff_put_vp8_epel8_h6v6_##opt; \
|
|
\
|
|
dsp->put_vp8_epel_pixels_tab[2][0][0] = ff_put_vp8_pixels4_armv6; \
|
|
dsp->put_vp8_epel_pixels_tab[2][0][1] = ff_put_vp8_epel4_h4_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[2][0][2] = ff_put_vp8_epel4_h6_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[2][1][0] = ff_put_vp8_epel4_v4_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[2][1][1] = ff_put_vp8_epel4_h4v4_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[2][1][2] = ff_put_vp8_epel4_h6v4_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[2][2][0] = ff_put_vp8_epel4_v6_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[2][2][1] = ff_put_vp8_epel4_h4v6_##opt; \
|
|
dsp->put_vp8_epel_pixels_tab[2][2][2] = ff_put_vp8_epel4_h6v6_##opt; \
|
|
\
|
|
dsp->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[0][0][2] = ff_put_vp8_bilin16_h_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[0][2][0] = ff_put_vp8_bilin16_v_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[0][2][2] = ff_put_vp8_bilin16_hv_##opt; \
|
|
\
|
|
dsp->put_vp8_bilinear_pixels_tab[1][0][0] = ff_put_vp8_pixels8_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[1][0][1] = ff_put_vp8_bilin8_h_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[1][0][2] = ff_put_vp8_bilin8_h_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[1][1][0] = ff_put_vp8_bilin8_v_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[1][1][1] = ff_put_vp8_bilin8_hv_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[1][1][2] = ff_put_vp8_bilin8_hv_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[1][2][0] = ff_put_vp8_bilin8_v_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[1][2][1] = ff_put_vp8_bilin8_hv_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[1][2][2] = ff_put_vp8_bilin8_hv_##opt; \
|
|
\
|
|
dsp->put_vp8_bilinear_pixels_tab[2][0][0] = ff_put_vp8_pixels4_armv6; \
|
|
dsp->put_vp8_bilinear_pixels_tab[2][0][1] = ff_put_vp8_bilin4_h_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[2][0][2] = ff_put_vp8_bilin4_h_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[2][1][0] = ff_put_vp8_bilin4_v_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[2][1][1] = ff_put_vp8_bilin4_hv_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[2][1][2] = ff_put_vp8_bilin4_hv_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[2][2][0] = ff_put_vp8_bilin4_v_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[2][2][1] = ff_put_vp8_bilin4_hv_##opt; \
|
|
dsp->put_vp8_bilinear_pixels_tab[2][2][2] = ff_put_vp8_bilin4_hv_##opt
|
|
if (HAVE_NEON) {
|
|
set_func_ptrs(neon);
|
|
} else if (HAVE_ARMV6) {
|
|
set_func_ptrs(armv6);
|
|
}
|
|
}
|