Skip to content
GitLab
Commit ac3d97f1 authored by Dmitry Kovalev's avatar Dmitry Kovalev
Browse files

Cleaning up vp9_variance_mmx.c. 

Change-Id: I42d83f91e272c92daed604c233f74439fe6307c5
parent 39b97318
v1.14.0-linphone 1.4.X feature/update_to_v1.9.0-linphone feature/uwp_nuget frame_parallel highbitdepth indianrunnerduck javanwhistlingduck khakicampbell linphone linphone-android linphone-old longtailedduck m49-2623 m52-2743 m54-2840 m56-2924 m66-3359 m68-3440 mandarinduck nextgen nextgenv2 playground sandbox/Jingning/experimental sandbox/Jingning/vpx sandbox/aconverse@google.com/ansbench sandbox/hkuang/frame_parallel sandbox/hkuang@google.com/decode sandbox/jimbankoski@google.com/proposed-aom sandbox/jingning@google.com/decoder_test_suite sandbox/jingning@google.com/experimental sandbox/jzern@google.com/test sandbox/wangch@google.com/vp9 sandbox/yaowu@google.com/mergeaom v1.12.0-linphone v1.6.1_linphone v1.7.0-linphone v1.9.0-linphone v1.9.0 v1.9.0-rc1 v1.8.2 v1.8.1 v1.8.0 v1.7.0 v1.6.1 v1.6.0 v1.5.0 v1.4.0
No related merge requests found
Hide whitespace changes
Inline Side-by-side
Showing
with 70 additions and 119 deletions
vp9/encoder/x86/vp9_variance_mmx.c
+ 70
− 119
View file @ ac3d97f1
...@@ -12,141 +12,92 @@ ...@@ -12,141 +12,92 @@
#include "vp9/encoder/vp9_variance.h" #include "vp9/encoder/vp9_variance.h"
#include "vpx_ports/mem.h" #include "vpx_ports/mem.h"
extern unsigned int vp9_get8x8var_mmx unsigned int vp9_get8x8var_mmx(const uint8_t *src, int src_stride,
( const uint8_t *ref, int ref_stride,
const unsigned char *src_ptr, unsigned int *sse, int *sum);
int source_stride,
const unsigned char *ref_ptr, unsigned int vp9_get4x4var_mmx(const uint8_t *src, int src_stride,
int recon_stride, const uint8_t *ref, int ref_stride,
unsigned int *SSE, unsigned int *SSE, int *sum);
int *Sum
); unsigned int vp9_variance4x4_mmx(const uint8_t *src, int src_stride,
extern unsigned int vp9_get4x4var_mmx const uint8_t *ref, int ref_stride,
( unsigned int *sse) {
const unsigned char *src_ptr, int sum;
int source_stride, vp9_get4x4var_mmx(src, src_stride, ref, ref_stride, sse, &sum);
const unsigned char *ref_ptr, return *sse - (((unsigned int)sum * sum) >> 4);
int recon_stride,
unsigned int *SSE,
int *Sum
);
unsigned int vp9_variance4x4_mmx(
const unsigned char *src_ptr,
int source_stride,
const unsigned char *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int var;
int avg;
vp9_get4x4var_mmx(src_ptr, source_stride, ref_ptr, recon_stride, &var, &avg);
*sse = var;
return (var - (((unsigned int)avg * avg) >> 4));
} }
unsigned int vp9_variance8x8_mmx( unsigned int vp9_variance8x8_mmx(const uint8_t *src, int src_stride,
const unsigned char *src_ptr, const uint8_t *ref, int ref_stride,
int source_stride, unsigned int *sse) {
const unsigned char *ref_ptr, int sum;
int recon_stride, vp9_get8x8var_mmx(src, src_stride, ref, ref_stride, sse, &sum);
unsigned int *sse) { return *sse - (((unsigned int)sum * sum) >> 6);
unsigned int var; }
int avg;
unsigned int vp9_mse16x16_mmx(const uint8_t *src, int src_stride,
const uint8_t *ref, int ref_stride,
unsigned int *sse) {
unsigned int sse0, sse1, sse2, sse3;
int sum0, sum1, sum2, sum3;
vp9_get8x8var_mmx(src_ptr, source_stride, ref_ptr, recon_stride, &var, &avg); vp9_get8x8var_mmx(src, src_stride, ref, ref_stride, &sse0, &sum0);
*sse = var; vp9_get8x8var_mmx(src + 8, src_stride, ref + 8, ref_stride, &sse1, &sum1);
vp9_get8x8var_mmx(src + 8 * src_stride, src_stride,
ref + 8 * ref_stride, ref_stride, &sse2, &sum2);
vp9_get8x8var_mmx(src + 8 * src_stride + 8, src_stride,
ref + 8 * ref_stride + 8, ref_stride, &sse3, &sum3);
return (var - (((unsigned int)avg * avg) >> 6)); *sse = sse0 + sse1 + sse2 + sse3;
return *sse;
} }
unsigned int vp9_mse16x16_mmx(
const unsigned char *src_ptr,
int source_stride,
const unsigned char *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int sse0, sse1, sse2, sse3, var;
int sum0, sum1, sum2, sum3;
unsigned int vp9_variance16x16_mmx(const uint8_t *src, int src_stride,
const uint8_t *ref, int ref_stride,
unsigned int *sse) {
unsigned int sse0, sse1, sse2, sse3;
int sum0, sum1, sum2, sum3, sum;
vp9_get8x8var_mmx(src_ptr, source_stride, ref_ptr, recon_stride, &sse0, vp9_get8x8var_mmx(src, src_stride, ref, ref_stride, &sse0, &sum0);
&sum0); vp9_get8x8var_mmx(src + 8, src_stride, ref + 8, ref_stride, &sse1, &sum1);
vp9_get8x8var_mmx(src_ptr + 8, source_stride, ref_ptr + 8, recon_stride, vp9_get8x8var_mmx(src + 8 * src_stride, src_stride,
&sse1, &sum1); ref + 8 * ref_stride, ref_stride, &sse2, &sum2);
vp9_get8x8var_mmx(src_ptr + 8 * source_stride, source_stride, vp9_get8x8var_mmx(src + 8 * src_stride + 8, src_stride,
ref_ptr + 8 * recon_stride, recon_stride, &sse2, &sum2); ref + 8 * ref_stride + 8, ref_stride, &sse3, &sum3);
vp9_get8x8var_mmx(src_ptr + 8 * source_stride + 8, source_stride,
ref_ptr + 8 * recon_stride + 8, recon_stride, &sse3, &sum3);
var = sse0 + sse1 + sse2 + sse3; *sse = sse0 + sse1 + sse2 + sse3;
*sse = var; sum = sum0 + sum1 + sum2 + sum3;
return var; return *sse - (((unsigned int)sum * sum) >> 8);
} }
unsigned int vp9_variance16x8_mmx(const uint8_t *src, int src_stride,
const uint8_t *ref, int ref_stride,
unsigned int *sse) {
unsigned int sse0, sse1;
int sum0, sum1, sum;
unsigned int vp9_variance16x16_mmx( vp9_get8x8var_mmx(src, src_stride, ref, ref_stride, &sse0, &sum0);
const unsigned char *src_ptr, vp9_get8x8var_mmx(src + 8, src_stride, ref + 8, ref_stride, &sse1, &sum1);
int source_stride,
const unsigned char *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int sse0, sse1, sse2, sse3, var;
int sum0, sum1, sum2, sum3, avg;
vp9_get8x8var_mmx(src_ptr, source_stride, ref_ptr, recon_stride, &sse0,
&sum0);
vp9_get8x8var_mmx(src_ptr + 8, source_stride, ref_ptr + 8, recon_stride,
&sse1, &sum1);
vp9_get8x8var_mmx(src_ptr + 8 * source_stride, source_stride,
ref_ptr + 8 * recon_stride, recon_stride, &sse2, &sum2);
vp9_get8x8var_mmx(src_ptr + 8 * source_stride + 8, source_stride,
ref_ptr + 8 * recon_stride + 8, recon_stride, &sse3, &sum3);
var = sse0 + sse1 + sse2 + sse3;
avg = sum0 + sum1 + sum2 + sum3;
*sse = var;
return (var - (((unsigned int)avg * avg) >> 8));
}
unsigned int vp9_variance16x8_mmx( *sse = sse0 + sse1;
const unsigned char *src_ptr, sum = sum0 + sum1;
int source_stride, return *sse - (((unsigned int)sum * sum) >> 7);
const unsigned char *ref_ptr,
int recon_stride,
unsigned int *sse) {
unsigned int sse0, sse1, var;
int sum0, sum1, avg;
vp9_get8x8var_mmx(src_ptr, source_stride, ref_ptr, recon_stride, &sse0,
&sum0);
vp9_get8x8var_mmx(src_ptr + 8, source_stride, ref_ptr + 8, recon_stride,
&sse1, &sum1);
var = sse0 + sse1;
avg = sum0 + sum1;
*sse = var;
return (var - (((unsigned int)avg * avg) >> 7));
} }
unsigned int vp9_variance8x16_mmx( unsigned int vp9_variance8x16_mmx(const uint8_t *src, int src_stride,
const unsigned char *src_ptr, const uint8_t *ref, int ref_stride,
int source_stride, unsigned int *sse) {
const unsigned char *ref_ptr, unsigned int sse0, sse1;
int recon_stride, int sum0, sum1, sum;
unsigned int *sse) {
unsigned int sse0, sse1, var;
int sum0, sum1, avg;
vp9_get8x8var_mmx(src_ptr, source_stride, ref_ptr, recon_stride, &sse0,
&sum0);
vp9_get8x8var_mmx(src_ptr + 8 * source_stride, source_stride,
ref_ptr + 8 * recon_stride, recon_stride, &sse1, &sum1);
var = sse0 + sse1; vp9_get8x8var_mmx(src, src_stride, ref, ref_stride, &sse0, &sum0);
avg = sum0 + sum1; vp9_get8x8var_mmx(src + 8 * src_stride, src_stride,
*sse = var; ref + 8 * ref_stride, ref_stride, &sse1, &sum1);
return (var - (((unsigned int)avg * avg) >> 7)); *sse = sse0 + sse1;
sum = sum0 + sum1;
return *sse - (((unsigned int)sum * sum) >> 7);
} }
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment