diff options
| author | Ronald S. Bultje <rsbultje@gmail.com> | 2010-07-19 21:18:04 +0000 |
|---|---|---|
| committer | Ronald S. Bultje <rsbultje@gmail.com> | 2010-07-19 21:18:04 +0000 |
| commit | 3facfc99daecd10c2b87761d111d4dee1e3736b7 (patch) | |
| tree | 987c1610b9ad1a6bf26f4181e26d91c8dbe842df /libavcodec/x86 | |
| parent | 581a96976b9061ed75b280e81fa1659ae3e395ee (diff) | |
| download | ffmpeg-3facfc99daecd10c2b87761d111d4dee1e3736b7.tar.gz | |
Change function prototypes for width=8 inner and mbedge loopfilter functions
so that it does both U and V planes at the same time. This will have speed
advantages when using SSE2 (or higher) optimizations, since we can do both
the U and V rows together in a single xmm register.
This also renames filter16 to filter16y and filter8 to filter8uv so that it's
more obvious what each function is used for.
Originally committed as revision 24337 to svn://svn.ffmpeg.org/ffmpeg/trunk
Diffstat (limited to 'libavcodec/x86')
| -rw-r--r-- | libavcodec/x86/vp8dsp-init.c | 36 | ||||
| -rw-r--r-- | libavcodec/x86/vp8dsp.asm | 2 |
2 files changed, 19 insertions, 19 deletions
diff --git a/libavcodec/x86/vp8dsp-init.c b/libavcodec/x86/vp8dsp-init.c index edff2883f7..dbaa0dc0a5 100644 --- a/libavcodec/x86/vp8dsp-init.c +++ b/libavcodec/x86/vp8dsp-init.c @@ -230,18 +230,18 @@ extern void ff_vp8_h_loop_filter_simple_mmx (uint8_t *dst, int stride, int fli extern void ff_vp8_h_loop_filter_simple_mmxext(uint8_t *dst, int stride, int flim); extern void ff_vp8_h_loop_filter_simple_sse2 (uint8_t *dst, int stride, int flim); -extern void ff_vp8_v_loop_filter16_inner_mmx (uint8_t *dst, int stride, - int e, int i, int hvt); -extern void ff_vp8_v_loop_filter16_inner_mmxext(uint8_t *dst, int stride, - int e, int i, int hvt); -extern void ff_vp8_v_loop_filter16_inner_sse2 (uint8_t *dst, int stride, - int e, int i, int hvt); -extern void ff_vp8_h_loop_filter16_inner_mmx (uint8_t *dst, int stride, - int e, int i, int hvt); -extern void ff_vp8_h_loop_filter16_inner_mmxext(uint8_t *dst, int stride, - int e, int i, int hvt); -extern void ff_vp8_h_loop_filter16_inner_sse2 (uint8_t *dst, int stride, - int e, int i, int hvt); +extern void ff_vp8_v_loop_filter16y_inner_mmx (uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_v_loop_filter16y_inner_mmxext(uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_v_loop_filter16y_inner_sse2 (uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_h_loop_filter16y_inner_mmx (uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_h_loop_filter16y_inner_mmxext(uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_h_loop_filter16y_inner_sse2 (uint8_t *dst, int stride, + int e, int i, int hvt); #endif #define VP8_LUMA_MC_FUNC(IDX, SIZE, OPT) \ @@ -284,8 +284,8 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmx; c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmx; - c->vp8_v_loop_filter16_inner = ff_vp8_v_loop_filter16_inner_mmx; - c->vp8_h_loop_filter16_inner = ff_vp8_h_loop_filter16_inner_mmx; + c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmx; + c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmx; } /* note that 4-tap width=16 functions are missing because w=16 @@ -302,8 +302,8 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmxext; c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmxext; - c->vp8_v_loop_filter16_inner = ff_vp8_v_loop_filter16_inner_mmxext; - c->vp8_h_loop_filter16_inner = ff_vp8_h_loop_filter16_inner_mmxext; + c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmxext; + c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmxext; } if (mm_flags & FF_MM_SSE) { @@ -320,8 +320,8 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2; c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse2; - c->vp8_v_loop_filter16_inner = ff_vp8_v_loop_filter16_inner_sse2; - c->vp8_h_loop_filter16_inner = ff_vp8_h_loop_filter16_inner_sse2; + c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2; + c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_sse2; } if (mm_flags & FF_MM_SSSE3) { diff --git a/libavcodec/x86/vp8dsp.asm b/libavcodec/x86/vp8dsp.asm index 70f0ae9286..116064c42f 100644 --- a/libavcodec/x86/vp8dsp.asm +++ b/libavcodec/x86/vp8dsp.asm @@ -1379,7 +1379,7 @@ SIMPLE_LOOPFILTER sse2, h, 6 ;----------------------------------------------------------------------------- %macro INNER_LOOPFILTER 4 -cglobal vp8_%2_loop_filter16_inner_%1, 5, %3, %4 +cglobal vp8_%2_loop_filter16y_inner_%1, 5, %3, %4 %define dst_reg r0 %define mstride_reg r1 %define E_reg r2 |