Implement sse2 and ssse3 versions for all sub_pixel_variance sizes.

Overall speedup around 5% (bus @ 1500kbps first 50 frames 4min10 ->
3min58). Specific changes to timings for each function compared to
original assembly-optimized versions (or just new version timings if
no previous assembly-optimized version was available):

sse2   4x4:    99 ->   82 cycles
sse2   4x8:           128 cycles
sse2   8x4:           121 cycles
sse2   8x8:   149 ->  129 cycles
sse2   8x16:  235 ->  245 cycles (?)
sse2  16x8:   269 ->  203 cycles
sse2  16x16:  441 ->  349 cycles
sse2  16x32:          641 cycles
sse2  32x16:          643 cycles
sse2  32x32: 1733 -> 1154 cycles
sse2  32x64:         2247 cycles
sse2  64x32:         2323 cycles
sse2  64x64: 6984 -> 4442 cycles

ssse3  4x4:           100 cycles (?)
ssse3  4x8:           103 cycles
ssse3  8x4:            71 cycles
ssse3  8x8:           147 cycles
ssse3  8x16:          158 cycles
ssse3 16x8:   188 ->  162 cycles
ssse3 16x16:  316 ->  273 cycles
ssse3 16x32:          535 cycles
ssse3 32x16:          564 cycles
ssse3 32x32:          973 cycles
ssse3 32x64:         1930 cycles
ssse3 64x32:         1922 cycles
ssse3 64x64:         3760 cycles

Change-Id: I81ff6fe51daf35a40d19785167004664d7e0c59d

1 files changed, 0 insertions, 235 deletions

diff --git a/vp9/encoder/x86/vp9_variance_mmx.c b/vp9/encoder/x86/vp9_variance_mmx.c
index bad1cfa74..d1415606e 100644
--- a/vp9/encoder/x86/vp9_variance_mmx.c
+++ b/vp9/encoder/x86/vp9_variance_mmx.c
@@ -13,27 +13,6 @@
 #include "vp9/common/vp9_pragmas.h"
 #include "vpx_ports/mem.h"
 
-extern void filter_block1d_h6_mmx
-(
-  const unsigned char *src_ptr,
-  unsigned short *output_ptr,
-  unsigned int src_pixels_per_line,
-  unsigned int pixel_step,
-  unsigned int output_height,
-  unsigned int output_width,
-  short *vp7_filter
-);
-extern void filter_block1d_v6_mmx
-(
-  const short *src_ptr,
-  unsigned char *output_ptr,
-  unsigned int pixels_per_line,
-  unsigned int pixel_step,
-  unsigned int output_height,
-  unsigned int output_width,
-  short *vp7_filter
-);
-
 extern unsigned int vp9_get_mb_ss_mmx(const short *src_ptr);
 extern unsigned int vp9_get8x8var_mmx
 (
@@ -53,30 +32,6 @@ extern unsigned int vp9_get4x4var_mmx
   unsigned int *SSE,
   int *Sum
 );
-extern void vp9_filter_block2d_bil4x4_var_mmx
-(
-  const unsigned char *ref_ptr,
-  int ref_pixels_per_line,
-  const unsigned char *src_ptr,
-  int src_pixels_per_line,
-  const short *HFilter,
-  const short *VFilter,
-  int *sum,
-  unsigned int *sumsquared
-);
-extern void vp9_filter_block2d_bil_var_mmx
-(
-  const unsigned char *ref_ptr,
-  int ref_pixels_per_line,
-  const unsigned char *src_ptr,
-  int src_pixels_per_line,
-  unsigned int Height,
-  const short *HFilter,
-  const short *VFilter,
-  int *sum,
-  unsigned int *sumsquared
-);
-
 
 unsigned int vp9_variance4x4_mmx(
   const unsigned char *src_ptr,
@@ -190,193 +145,3 @@ unsigned int vp9_variance8x16_mmx(
   return (var - (((unsigned int)avg * avg) >> 7));
 
 }
-
-DECLARE_ALIGNED(16, extern const short, vp9_bilinear_filters_mmx[16][8]);
-
-unsigned int vp9_sub_pixel_variance4x4_mmx
-(
-  const unsigned char  *src_ptr,
-  int  src_pixels_per_line,
-  int  xoffset,
-  int  yoffset,
-  const unsigned char *dst_ptr,
-  int dst_pixels_per_line,
-  unsigned int *sse)
-
-{
-  int xsum;
-  unsigned int xxsum;
-  vp9_filter_block2d_bil4x4_var_mmx(
-    src_ptr, src_pixels_per_line,
-    dst_ptr, dst_pixels_per_line,
-    vp9_bilinear_filters_mmx[xoffset], vp9_bilinear_filters_mmx[yoffset],
-    &xsum, &xxsum
-  );
-  *sse = xxsum;
-  return (xxsum - (((unsigned int)xsum * xsum) >> 4));
-}
-
-
-unsigned int vp9_sub_pixel_variance8x8_mmx
-(
-  const unsigned char  *src_ptr,
-  int  src_pixels_per_line,
-  int  xoffset,
-  int  yoffset,
-  const unsigned char *dst_ptr,
-  int dst_pixels_per_line,
-  unsigned int *sse
-) {
-
-  int xsum;
-  unsigned int xxsum;
-  vp9_filter_block2d_bil_var_mmx(
-    src_ptr, src_pixels_per_line,
-    dst_ptr, dst_pixels_per_line, 8,
-    vp9_bilinear_filters_mmx[xoffset], vp9_bilinear_filters_mmx[yoffset],
-    &xsum, &xxsum
-  );
-  *sse = xxsum;
-  return (xxsum - (((unsigned int)xsum * xsum) >> 6));
-}
-
-unsigned int vp9_sub_pixel_variance16x16_mmx
-(
-  const unsigned char  *src_ptr,
-  int  src_pixels_per_line,
-  int  xoffset,
-  int  yoffset,
-  const unsigned char *dst_ptr,
-  int dst_pixels_per_line,
-  unsigned int *sse
-) {
-
-  int xsum0, xsum1;
-  unsigned int xxsum0, xxsum1;
-
-  vp9_filter_block2d_bil_var_mmx(
-    src_ptr, src_pixels_per_line,
-    dst_ptr, dst_pixels_per_line, 16,
-    vp9_bilinear_filters_mmx[xoffset], vp9_bilinear_filters_mmx[yoffset],
-    &xsum0, &xxsum0
-  );
-
-  vp9_filter_block2d_bil_var_mmx(
-    src_ptr + 8, src_pixels_per_line,
-    dst_ptr + 8, dst_pixels_per_line, 16,
-    vp9_bilinear_filters_mmx[xoffset], vp9_bilinear_filters_mmx[yoffset],
-    &xsum1, &xxsum1
-  );
-
-  xsum0 += xsum1;
-  xxsum0 += xxsum1;
-
-  *sse = xxsum0;
-  return (xxsum0 - (((unsigned int)xsum0 * xsum0) >> 8));
-
-
-}
-
-unsigned int vp9_sub_pixel_mse16x16_mmx(
-  const unsigned char  *src_ptr,
-  int  src_pixels_per_line,
-  int  xoffset,
-  int  yoffset,
-  const unsigned char *dst_ptr,
-  int dst_pixels_per_line,
-  unsigned int *sse
-) {
-  vp9_sub_pixel_variance16x16_mmx(src_ptr, src_pixels_per_line, xoffset, yoffset, dst_ptr, dst_pixels_per_line, sse);
-  return *sse;
-}
-
-unsigned int vp9_sub_pixel_variance16x8_mmx
-(
-  const unsigned char  *src_ptr,
-  int  src_pixels_per_line,
-  int  xoffset,
-  int  yoffset,
-  const unsigned char *dst_ptr,
-  int dst_pixels_per_line,
-  unsigned int *sse
-) {
-  int xsum0, xsum1;
-  unsigned int xxsum0, xxsum1;
-
-
-  vp9_filter_block2d_bil_var_mmx(
-    src_ptr, src_pixels_per_line,
-    dst_ptr, dst_pixels_per_line, 8,
-    vp9_bilinear_filters_mmx[xoffset], vp9_bilinear_filters_mmx[yoffset],
-    &xsum0, &xxsum0
-  );
-
-
-  vp9_filter_block2d_bil_var_mmx(
-    src_ptr + 8, src_pixels_per_line,
-    dst_ptr + 8, dst_pixels_per_line, 8,
-    vp9_bilinear_filters_mmx[xoffset], vp9_bilinear_filters_mmx[yoffset],
-    &xsum1, &xxsum1
-  );
-
-  xsum0 += xsum1;
-  xxsum0 += xxsum1;
-
-  *sse = xxsum0;
-  return (xxsum0 - (((unsigned int)xsum0 * xsum0) >> 7));
-}
-
-unsigned int vp9_sub_pixel_variance8x16_mmx
-(
-  const unsigned char  *src_ptr,
-  int  src_pixels_per_line,
-  int  xoffset,
-  int  yoffset,
-  const unsigned char *dst_ptr,
-  int dst_pixels_per_line,
-  unsigned int *sse
-) {
-  int xsum;
-  unsigned int xxsum;
-  vp9_filter_block2d_bil_var_mmx(
-    src_ptr, src_pixels_per_line,
-    dst_ptr, dst_pixels_per_line, 16,
-    vp9_bilinear_filters_mmx[xoffset], vp9_bilinear_filters_mmx[yoffset],
-    &xsum, &xxsum
-  );
-  *sse = xxsum;
-  return (xxsum - (((unsigned int)xsum * xsum) >> 7));
-}
-
-
-unsigned int vp9_variance_halfpixvar16x16_h_mmx(
-  const unsigned char *src_ptr,
-  int  source_stride,
-  const unsigned char *ref_ptr,
-  int  recon_stride,
-  unsigned int *sse) {
-  return vp9_sub_pixel_variance16x16_mmx(src_ptr, source_stride, 8, 0,
-                                         ref_ptr, recon_stride, sse);
-}
-
-
-unsigned int vp9_variance_halfpixvar16x16_v_mmx(
-  const unsigned char *src_ptr,
-  int  source_stride,
-  const unsigned char *ref_ptr,
-  int  recon_stride,
-  unsigned int *sse) {
-  return vp9_sub_pixel_variance16x16_mmx(src_ptr, source_stride, 0, 8,
-                                         ref_ptr, recon_stride, sse);
-}
-
-
-unsigned int vp9_variance_halfpixvar16x16_hv_mmx(
-  const unsigned char *src_ptr,
-  int  source_stride,
-  const unsigned char *ref_ptr,
-  int  recon_stride,
-  unsigned int *sse) {
-  return vp9_sub_pixel_variance16x16_mmx(src_ptr, source_stride, 8, 8,
-                                         ref_ptr, recon_stride, sse);
-}