Hi,

there are numerous places with SSSE3 code using padd(s)w+psraw
instructions instead of the pmulhrsw one, which adds about 1 cycle per
iteration.

vp8 MC is one of them.

Christophe

Attachment: 0003-vp8dsp-x86-perform-rounding-shift-with-a-single-inst.patch
Description: Binary data

_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to