]> granicus.if.org Git - libvpx/commit
projects / libvpx / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: a2b598a) | patch
Add SSE2 subtract functions
authorYunqing Wang <yunqingwang@google.com>
Mon, 18 Oct 2010 18:15:15 +0000 (14:15 -0400)
committerYunqing Wang <yunqingwang@google.com>
Mon, 18 Oct 2010 18:15:15 +0000 (14:15 -0400)
commit4db2076594be3a48c6c1b3755c1d9621f5ad1c5b
tree831bdf348995694e320a7520f6f5a4804a37f964tree | snapshot
parenta2b598a2f964b0d7a1e5df74b66c70a1bca746cacommit | diff
Add SSE2 subtract functions

Instead of doing 8-bit data unpack and 16-bit subtraction, use
psubb to do 16 8-bit subtractions and pcmpgtb to preserve the
sign information. This does not bring noticable gain since
these functions are not called frequently.

Change-Id: I90a0dfaa3db9d422e4ada324076596ffb178548e
vp8/encoder/x86/encodemb_x86.h diff | blob | history
vp8/encoder/x86/subtract_mmx.asm diff | blob | history
vp8/encoder/x86/subtract_sse2.asm [new file with mode: 0644] blob
vp8/encoder/x86/x86_csystemdependent.c diff | blob | history
vp8/vp8cx.mk diff | blob | history
Unnamed repository; edit this file 'description' to name the repository.