| Fri, 15 Feb 2008 21:52:34 +0000 |
uau |
libvo: change asm syntax to use ASMALIGN and " # nop"
|
| Sun, 27 May 2007 21:45:33 +0000 |
michael |
HAVE_MMX1 -> HAVE_ONLY_MMX1 (makes more sense ...)
|
| Sat, 04 Jun 2005 21:11:39 +0000 |
gpoirier |
More gcc-4.0 fixes
|
| Fri, 21 Jan 2005 21:11:35 +0000 |
gpoirier |
Make this file compile with gcc-4.0.0. The old code was invalid C.
|
| Thu, 21 Oct 2004 11:55:20 +0000 |
aurel |
adapting existing mmx/mmx2/sse/3dnow optimizations so they work on x86_64
|
| Thu, 22 Aug 2002 23:28:33 +0000 |
arpi |
removed nonsense 'inline'
|
| Wed, 17 Apr 2002 19:11:52 +0000 |
michael |
prefer prefetchnta if its available
|
| Wed, 17 Apr 2002 18:48:22 +0000 |
michael |
pre mmx2/3dnow fix
|
| Wed, 13 Feb 2002 00:17:16 +0000 |
michael |
101
|
| Tue, 12 Feb 2002 23:17:14 +0000 |
michael |
mem2agpcpy()
|
| Sun, 09 Dec 2001 15:25:11 +0000 |
michael |
runtime cpu detection
|
| Thu, 22 Nov 2001 19:40:38 +0000 |
michael |
10-20% faster fastmemcpy :) on my p3 at least but the algo is mostly from "amd athlon processor x86 code optimization guide" so it should be faster for amd chips too, but i fear it might be slower for mem->vram copies (someone should check that, i cant) ... there are 2 #defines to finetune it (BLOCK_SIZE & CONFUSION_FACTOR)
|
| Wed, 13 Jun 2001 16:12:14 +0000 |
nickols_k |
Issues about P3 performance and SSE2 support.
|
| Sat, 05 May 2001 13:30:00 +0000 |
nickols_k |
Moving fast_memcpy to separate file (Size optimization)
|
