[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] OSX SDL altivec optimization
From: |
Magnus Damm |
Subject: |
Re: [Qemu-devel] OSX SDL altivec optimization |
Date: |
Thu, 27 Jan 2005 23:37:56 +0100 |
On Thu, 27 Jan 2005 23:20:33 +0100, Laurent Amon
<address@hidden> wrote:
> For those of you who have a G4, I just had a look on the SDL mailing
> lists archives and the following patch (in progress)
> http://twomix.devolution.com/pipermail/sdl/2005-January/066678.html
> adds altivec use to some blitting functions.
>
> I just recompiled 1.2.8 with -faltivec -mcpu=7400 -O3 and this patch
> and the drawing feels somewhat faster. Can somebody else confirm this?
Speaking about Altivec, ppc users running Linux with a gcc from FSF
should use "-mcpu=7400 -maltivec -mabi=altivec" instead. Altivec is
great and everything, but the history of Apple vs FSF gcc with flags,
include files, register names, macros and bugs is a big PITA. Have a
look at the configure script shipped with mplayer. There are actually
versions of Apple gcc that accepts but silently ignores the altivec
flags used by FSF gcc... Welcome to hell.
But what I was supposed to say was: Yay Altivec! =)
/ magnus