First shot at VP8 optimizations:

- MMXEXT, SSE2 and SSSE3 MC functions
- MMX and SSE4 IDCT dc_add functions

Patch by Jason Garrett-Glaser <darkshikari gmail com> and myself.

Originally committed as revision 23815 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavcodec/vp8dsp.c

@@ -451,4 +451,7 @@ av_cold void ff_vp8dsp_init(VP8DSPContext *dsp)
     VP8_BILINEAR_MC_FUNC(0, 16);
     VP8_BILINEAR_MC_FUNC(1, 8);
     VP8_BILINEAR_MC_FUNC(2, 4);
+
+    if (ARCH_X86)
+        ff_vp8dsp_init_x86(dsp);
 }

libavcodec/vp8dsp.h

@@ -66,5 +66,6 @@ void ff_put_vp8_pixels8_c(uint8_t *dst, uint8_t *src, int stride, int h, int x,
 void ff_put_vp8_pixels4_c(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y);
 
 void ff_vp8dsp_init(VP8DSPContext *c);
+void ff_vp8dsp_init_x86(VP8DSPContext *c);
 
 #endif /* AVCODEC_VP8DSP_H */

libavcodec/x86/Makefile

@@ -30,6 +30,8 @@ MMX-OBJS-$(CONFIG_VP6_DECODER)         += x86/vp3dsp_mmx.o              \
                                           x86/vp3dsp_sse2.o             \
                                           x86/vp6dsp_mmx.o              \
                                           x86/vp6dsp_sse2.o
+YASM-OBJS-$(CONFIG_VP8_DECODER)        += x86/vp8dsp.o
+MMX-OBJS-$(CONFIG_VP8_DECODER)         += x86/vp8dsp-init.o
 MMX-OBJS-$(HAVE_YASM)                  += x86/dsputil_yasm.o            \
                                           $(YASM-OBJS-yes)