Log of /branches/dev-api-4/xvidcore/src/motion/x86_asm
Directory Listing
Revision
1205 -
Directory Listing
Modified
Thu Nov 13 23:11:24 2003 UTC (20 years, 5 months ago) by
edgomez
MMXed the calculation of SSE for 8x8 16bit blocks. This helps quite
a lot VHQ=4 mode.
My tests show with trellis:chroma_me:
- ~20% speed improvement for vhq=4.
- at least 5% when using vhq=1.
Of course this speedup vanishes if more CPU intensive features are
used. CruNcher who used gmc/qpel, noticed "only" a ~5% speed
improvement.
NB: i'm of course talking about overall speed improvement. Such a
small patch for such a big improvement :-)
Revision
1199 -
Directory Listing
Modified
Mon Nov 3 19:58:16 2003 UTC (20 years, 6 months ago) by
edgomez
* Small error fixed by Skal in his dev16 code (missing pshufd).
* Blocks used by DCT tests are now aligned with DECLARE_ALIGNED_MATRIX
this avoids the well know segfaults when using SSE2 instructions that
suppose data alignment.
Revision
1198 -
Directory Listing
Modified
Mon Nov 3 15:51:50 2003 UTC (20 years, 6 months ago) by
edgomez
correct .rodata alignment
Revision
1192 -
Directory Listing
Modified
Tue Oct 28 22:23:03 2003 UTC (20 years, 6 months ago) by
edgomez
* Applied same style to all asm files
* Replaced current sad sse2 operators with skal's ones
* Removed old and unused colorspace asm files
Revision
886 -
Directory Listing
Modified
Fri Feb 21 14:49:29 2003 UTC (21 years, 2 months ago) by
This commit was manufactured by cvs2svn to create branch 'dev-api-4'.