Merge "mips msa vp9 subpel variance optimization"

1 files changed, 13 insertions, 13 deletions

diff --git a/vp9/common/vp9_rtcd_defs.pl b/vp9/common/vp9_rtcd_defs.pl
index e3ab06d93..ec7da5c0c 100644
--- a/vp9/common/vp9_rtcd_defs.pl
+++ b/vp9/common/vp9_rtcd_defs.pl
@@ -805,80 +805,80 @@ if (vpx_config("CONFIG_VP9_ENCODER") eq "yes") {
 
 # variance
 add_proto qw/unsigned int vp9_sub_pixel_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance64x64 avx2 neon/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance64x64 avx2 neon msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance64x64 avx2/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance32x64/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance32x64 msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance32x64/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance64x32/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance64x32 msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance64x32/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance32x16/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance32x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance32x16/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance16x32/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance16x32 msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance16x32/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance32x32 avx2 neon/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance32x32 avx2 neon msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance32x32 avx2/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance16x16 neon/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance16x16 neon msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance16x16/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance8x16/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance8x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance8x16/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance16x8/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance16x8 msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance16x8/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance8x8 neon/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance8x8 neon msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance8x8/, "$sse2_x86inc", "$ssse3_x86inc";
 
 # TODO(jingning): need to convert 8x4/4x8 functions into mmx/sse form
 add_proto qw/unsigned int vp9_sub_pixel_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance8x4/, "$sse2_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance8x4 msa/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance8x4/, "$sse2_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance4x8/, "$sse_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance4x8 msa/, "$sse_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";
 specialize qw/vp9_sub_pixel_avg_variance4x8/, "$sse_x86inc", "$ssse3_x86inc";
 
 add_proto qw/unsigned int vp9_sub_pixel_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse";
-specialize qw/vp9_sub_pixel_variance4x4/, "$sse_x86inc", "$ssse3_x86inc";
+specialize qw/vp9_sub_pixel_variance4x4 msa/, "$sse_x86inc", "$ssse3_x86inc";
 #vp9_sub_pixel_variance4x4_sse2=vp9_sub_pixel_variance4x4_wmt
 
 add_proto qw/unsigned int vp9_sub_pixel_avg_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred";