generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Fix the reading of too many input pixels

Browse Source 
in VP9_get4x4var_mmx

Change-Id: I4b4a8f45f25ebdfad281f169cc87aba5e2d6f227
This commit is contained in:
Yaowu Xu 2013-08-26 09:21:12 -07:00
parent 66ccf5ddcf
commit 6c5433c836
1 changed files with 8 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
vp9/encoder/x86/vp9_variance_impl_mmx.asm
View File

@ -342,8 +342,8 @@ sym(vp9_get4x4var_mmx):
movsxd rdx, dword ptr arg(3) ;[recon_stride]
; Row 1
movq mm0, [rax] ; Copy eight bytes to mm0
movq mm1, [rbx] ; Copy eight bytes to mm1
movd mm0, [rax] ; Copy 4 bytes to mm0
movd mm1, [rbx] ; Copy 4 bytes to mm1
punpcklbw mm0, mm6 ; unpack to higher prrcision
punpcklbw mm1, mm6
psubsw mm0, mm1 ; A-B (low order) to MM0
@ -351,12 +351,12 @@ sym(vp9_get4x4var_mmx):
pmaddwd mm0, mm0 ; square and accumulate
add rbx,rdx ; Inc pointer into ref data
add rax,rcx ; Inc pointer into the new data
movq mm1, [rbx] ; Copy eight bytes to mm1
movd mm1, [rbx] ; Copy 4 bytes to mm1
paddd mm7, mm0 ; accumulate in mm7
; Row 2
movq mm0, [rax] ; Copy eight bytes to mm0
movd mm0, [rax] ; Copy 4 bytes to mm0
punpcklbw mm0, mm6 ; unpack to higher prrcision
punpcklbw mm1, mm6
psubsw mm0, mm1 ; A-B (low order) to MM0
@ -365,11 +365,11 @@ sym(vp9_get4x4var_mmx):
pmaddwd mm0, mm0 ; square and accumulate
add rbx,rdx ; Inc pointer into ref data
add rax,rcx ; Inc pointer into the new data
movq mm1, [rbx] ; Copy eight bytes to mm1
movd mm1, [rbx] ; Copy 4 bytes to mm1
paddd mm7, mm0 ; accumulate in mm7
; Row 3
movq mm0, [rax] ; Copy eight bytes to mm0
movd mm0, [rax] ; Copy 4 bytes to mm0
punpcklbw mm0, mm6 ; unpack to higher prrcision
punpcklbw mm1, mm6
psubsw mm0, mm1 ; A-B (low order) to MM0
@ -378,11 +378,11 @@ sym(vp9_get4x4var_mmx):
pmaddwd mm0, mm0 ; square and accumulate
add rbx,rdx ; Inc pointer into ref data
add rax,rcx ; Inc pointer into the new data
movq mm1, [rbx] ; Copy eight bytes to mm1
movd mm1, [rbx] ; Copy 4 bytes to mm1
paddd mm7, mm0 ; accumulate in mm7
; Row 4
movq mm0, [rax] ; Copy eight bytes to mm0
movd mm0, [rax] ; Copy 4 bytes to mm0
punpcklbw mm0, mm6 ; unpack to higher prrcision
punpcklbw mm1, mm6