apps/ffmpeg-rockchip
1
0
Fork 0
mirror of https://github.com/nyanmisaka/ffmpeg-rockchip.git synced 2025-11-01 04:53:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

VP8: optimize DC-only chroma case in the same way as luma.

Browse Source 
Add MMX idct_dc_add4uv function for this case.
~40% faster chroma idct.

Originally committed as revision 24455 to svn://svn.ffmpeg.org/ffmpeg/trunk
This commit is contained in:
Jason Garrett-Glaser
2010-07-23 06:02:52 +00:00
parent 3df56f4118
commit 3ae079a3c8
5 changed files with 89 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
libavcodec/x86/vp8dsp-init.c
View File

@@ -220,8 +220,9 @@ HVBILIN(ssse3, 8, 16, 16)
extern void ff_vp8_idct_dc_add_mmx(uint8_t *dst, DCTELEM block[16], int stride);
extern void ff_vp8_idct_dc_add_sse4(uint8_t *dst, DCTELEM block[16], int stride);
extern void ff_vp8_idct_dc_add4_mmx(uint8_t *dst, DCTELEM block[4][16], int stride);
extern void ff_vp8_idct_dc_add4_sse2(uint8_t *dst, DCTELEM block[4][16], int stride);
extern void ff_vp8_idct_dc_add4y_mmx(uint8_t *dst, DCTELEM block[4][16], int stride);
extern void ff_vp8_idct_dc_add4y_sse2(uint8_t *dst, DCTELEM block[4][16], int stride);
extern void ff_vp8_idct_dc_add4uv_mmx(uint8_t *dst, DCTELEM block[2][16], int stride);
extern void ff_vp8_luma_dc_wht_mmx(DCTELEM block[4][4][16], DCTELEM dc[16]);
extern void ff_vp8_idct_add_mmx(uint8_t *dst, DCTELEM block[16], int stride);
extern void ff_vp8_idct_add_sse(uint8_t *dst, DCTELEM block[16], int stride);
@@ -284,10 +285,11 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c)
#if HAVE_YASM
if (mm_flags & FF_MM_MMX) {
c->vp8_idct_dc_add = ff_vp8_idct_dc_add_mmx;
c->vp8_idct_dc_add4 = ff_vp8_idct_dc_add4_mmx;
c->vp8_idct_add = ff_vp8_idct_add_mmx;
c->vp8_luma_dc_wht = ff_vp8_luma_dc_wht_mmx;
c->vp8_idct_dc_add = ff_vp8_idct_dc_add_mmx;
c->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_mmx;
c->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_mmx;
c->vp8_idct_add = ff_vp8_idct_add_mmx;
c->vp8_luma_dc_wht = ff_vp8_luma_dc_wht_mmx;
c->put_vp8_epel_pixels_tab[0][0][0] =
c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_mmx;
c->put_vp8_epel_pixels_tab[1][0][0] =
@@ -354,7 +356,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c)
}
if (mm_flags & FF_MM_SSE2) {
c->vp8_idct_dc_add4 = ff_vp8_idct_dc_add4_sse2;
c->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_sse2;
c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_sse2;
c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_sse2;