ffmpeg/FFmpeg
1
0
Fork 0
mirror of https://git.ffmpeg.org/ffmpeg.git synced 2024-10-20 15:13:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

x86/aacpsdsp: optimize ff_ps_stereo_interpolate_sse3

Browse Source 
Move the unpacking outside of the loop. 5% to 10% faster.

Suggested-by: ubitux
Signed-off-by: James Almer <jamrial@gmail.com>
This commit is contained in:
James Almer 2017-06-02 19:17:28 -03:00
parent 2ba896fef7
commit be3809a521
1 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
libavcodec/x86/aacpsdsp.asm
View File

@ -93,6 +93,10 @@ cglobal ps_stereo_interpolate, 5, 5, 6, l, r, h, h_step, n
movaps m1, [h_stepq]
cmp nd, 0
jle .ret
unpcklps m4, m0, m0
unpckhps m0, m0
unpcklps m5, m1, m1
unpckhps m1, m1
shl nd, 3
add lq, nq
add rq, nq
@ -100,15 +104,12 @@ cglobal ps_stereo_interpolate, 5, 5, 6, l, r, h, h_step, n
align 16
.loop:
addps m4, m5
addps m0, m1
movddup m2, [lq+nq]
movddup m3, [rq+nq]
movaps m4, m0
movaps m5, m0
unpcklps m4, m4
unpckhps m5, m5
mulps m2, m4
mulps m3, m5
mulps m3, m0
addps m2, m3
movsd [lq+nq], m2
movhps [rq+nq], m2