Mercurial > libavcodec.hg
diff x86/vp8dsp-init.c @ 12205:d38e8565ba05 libavcodec
VP8 MBedge loopfilter MMX/MMX2/SSE2 functions for both luma (width=16)
and chroma (width=8).
| author | rbultje |
|---|---|
| date | Tue, 20 Jul 2010 22:58:56 +0000 |
| parents | 563339ea87aa |
| children | 9eef00a43280 |
line wrap: on
line diff
--- a/x86/vp8dsp-init.c Tue Jul 20 22:04:18 2010 +0000 +++ b/x86/vp8dsp-init.c Tue Jul 20 22:58:56 2010 +0000 @@ -255,6 +255,32 @@ int s, int e, int i, int hvt); extern void ff_vp8_h_loop_filter8uv_inner_sse2 (uint8_t *dstU, uint8_t *dstV, int s, int e, int i, int hvt); + +extern void ff_vp8_v_loop_filter16y_mbedge_mmx (uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_v_loop_filter16y_mbedge_mmxext(uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_v_loop_filter16y_mbedge_sse2 (uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_h_loop_filter16y_mbedge_mmx (uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_h_loop_filter16y_mbedge_mmxext(uint8_t *dst, int stride, + int e, int i, int hvt); +extern void ff_vp8_h_loop_filter16y_mbedge_sse2 (uint8_t *dst, int stride, + int e, int i, int hvt); + +extern void ff_vp8_v_loop_filter8uv_mbedge_mmx (uint8_t *dstU, uint8_t *dstV, + int s, int e, int i, int hvt); +extern void ff_vp8_v_loop_filter8uv_mbedge_mmxext(uint8_t *dstU, uint8_t *dstV, + int s, int e, int i, int hvt); +extern void ff_vp8_v_loop_filter8uv_mbedge_sse2 (uint8_t *dstU, uint8_t *dstV, + int s, int e, int i, int hvt); +extern void ff_vp8_h_loop_filter8uv_mbedge_mmx (uint8_t *dstU, uint8_t *dstV, + int s, int e, int i, int hvt); +extern void ff_vp8_h_loop_filter8uv_mbedge_mmxext(uint8_t *dstU, uint8_t *dstV, + int s, int e, int i, int hvt); +extern void ff_vp8_h_loop_filter8uv_mbedge_sse2 (uint8_t *dstU, uint8_t *dstV, + int s, int e, int i, int hvt); #endif #define VP8_LUMA_MC_FUNC(IDX, SIZE, OPT) \ @@ -301,6 +327,11 @@ c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmx; c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmx; c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmx; + + c->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16y_mbedge_mmx; + c->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16y_mbedge_mmx; + c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_mmx; + c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_mmx; } /* note that 4-tap width=16 functions are missing because w=16 @@ -321,6 +352,11 @@ c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmxext; c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmxext; c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmxext; + + c->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16y_mbedge_mmxext; + c->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16y_mbedge_mmxext; + c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_mmxext; + c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_mmxext; } if (mm_flags & FF_MM_SSE) { @@ -339,11 +375,17 @@ c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2; c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_sse2; + + c->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16y_mbedge_mmxext; + c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_mmxext; } if (mm_flags & FF_MM_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; + + c->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16y_mbedge_mmxext; + c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_mmxext; } if (mm_flags & FF_MM_SSSE3) {