Public SourceMage Discussion List

[Andrew <afrayedknot AT thefrayedknot.armory.com>]

On Mon, Apr 04, 2005 at 09:35:51PM -0700, Eric Sandall wrote:
> Quoting "David Michael Leo Brown Jr." <dmlb2000 AT excite.com>:
> > I was looking at the x86-64 specfile and was surprised to see that it 
> > wasn't
> > x86-64 the -march flag was set to k8 for gcc 3.4, when it should be 
> > x86-64.
> >
> > I understand that there's some people in sourcemage who have amd64 
> > processors
> > however, I don't, I have the intel version that uses the x86-64
> > specifications
> > (called extended memory 64 technology)
> >
> > gcc doesn't have em64t as a valid arch spec (not even in v4.1) the closest
> > thing I
> > can pass to gcc is x86-64
> >
> > I would propose a change to the specfiles and move the current x86-64 and
> > x86-32
> > to k8-64 and k8-32, and change the x86-64 and x86-32 to the more generic
> > -march=x86-64, also I would change the x86-[64,32] gcc 3.3 to 
> > -march=pentium4
> >
> > This way would provide arch specs for both amd and intel implementations 
> > of
> > the
> > arch specification x86-64.
> >
> > -- David Brown
> 
> Why not make archspecs for your architecture instead of modifying someone 
> else's
> to work on yours?
> 

I think the problem is, the x86_64 archspecs are false advertising, they
technically apply to k8 processors, not gcc's definition of x86-64. The
x86_64 archspec appears to have been incorrectly named.

-Andrew

-- 
__________________________________________________________________________
|Andrew D. Stitt                           | astitt at sourcemage.org    |
|irc: afrayedknot                          | afrayedknot at t.armory.com |
|aim: thefrayedknot or iteratorplusplus    | acedit at armory.com        |
|Sorcery Team Lead                         | ftp://t.armory.com/         |
--------------------------------------------------------------------------

Attachment: pgp5V9Bh4NrC4.pgp
Description: PGP signature