[gentoo-dev] The SIMD debacle

[gentoo-dev] The SIMD debacle

[Jon Ellis]

Can someone explain the sse / 3dnow! situation to me?

My understanding is that most modern intel and amd chips supprt sse, 
but only amd support 3dnow! Is this correct? Why would someone with a 
pentium 4 box have 3dnow in their use flags?

Does anyone point me a ebuilds that correctly handle this stuff? It 
would be most useful to understand how protect users from themselves 
and reduce the number 'invalid instruction' bugs filed against 
libvorbis!

Thanks

j.


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] The SIMD debacle

[Paul de Vrieze]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 30 September 2003 14:46, Jon Ellis wrote:
> Can someone explain the sse / 3dnow! situation to me?
>
> My understanding is that most modern intel and amd chips supprt sse,
> but only amd support 3dnow! Is this correct? Why would someone with a
> pentium 4 box have 3dnow in their use flags?
>

I think you are completely right, 3dnow in the useflags should be only done on 
amd systems that have 3dnow.

> Does anyone point me a ebuilds that correctly handle this stuff? It
> would be most useful to understand how protect users from themselves
> and reduce the number 'invalid instruction' bugs filed against
> libvorbis!

I think you might want to put in a check to test for actual 3dnow support. 
However it is perfectly legal to build a package on a p4 that is going to run 
on an athlon, so disabling the useflag might limit the users.

Paul

- -- 
Paul de Vrieze
Gentoo Developer
Mail: pauldv@gentoo.org
Homepage: http://www.devrieze.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/eX2mbKx5DBjWFdsRAlriAKCNN/lgHcVFYdZaceVvD5Kt0sgWAgCgiqgi
0zGRSJKQOS0stKTWREmM5II=
=Hv0F
-----END PGP SIGNATURE-----


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] The SIMD debacle

[Jon Portnoy]

On Tue, Sep 30, 2003 at 09:46:49PM +0900, Jon Ellis wrote:
> Can someone explain the sse / 3dnow! situation to me?
> 
> My understanding is that most modern intel and amd chips supprt sse, 
> but only amd support 3dnow! Is this correct? Why would someone with a 
> pentium 4 box have 3dnow in their use flags?
> 

They shouldn't. It was, however, in make.defaults at one point - I have 
no idea why.

I removed it a while back.

-- 
Jon Portnoy
avenj/irc.freenode.net

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] The SIMD debacle

[Chris Gianelloni]

[-- Attachment #1: Type: text/plain, Size: 939 bytes --]

On Tue, 2003-09-30 at 08:46, Jon Ellis wrote:
> Can someone explain the sse / 3dnow! situation to me?
> 
> My understanding is that most modern intel and amd chips supprt sse, 
> but only amd support 3dnow! Is this correct? Why would someone with a 
> pentium 4 box have 3dnow in their use flags?

Pentium 3 - sse
Pentium 4 - sse/sse2
Athlon - 3dnow
AthlonXP - 3dnow/sse

> Does anyone point me a ebuilds that correctly handle this stuff? It 
> would be most useful to understand how protect users from themselves 
> and reduce the number 'invalid instruction' bugs filed against 
> libvorbis!

If you need some help with this ebuild, I'll gladly help, since I think
I might have been the one that caused you all the headache in the first
place.

> 
> Thanks
> 
> j.
> 
> 
> --
> gentoo-dev@gentoo.org mailing list
-- 
Chris Gianelloni
Developer, Gentoo Linux
Games Team

Is your power animal a pengiun?

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] The SIMD debacle

[David Holm]

[-- Attachment #1: Type: text/plain, Size: 525 bytes --]

On Tue, 30 Sep 2003 12:21:26 -0400
Chris Gianelloni <wolf31o2@gentoo.org> wrote:

> On Tue, 2003-09-30 at 08:46, Jon Ellis wrote:
> > Can someone explain the sse / 3dnow! situation to me?
> > 
> > My understanding is that most modern intel and amd chips supprt sse, 
> > but only amd support 3dnow! Is this correct? Why would someone with a 
> > pentium 4 box have 3dnow in their use flags?
> 
> Pentium 3 - sse
> Pentium 4 - sse/sse2
> Athlon - 3dnow
> AthlonXP - 3dnow/sse

PPC74X0 - AltiVec
PPC970 - AltiVec

//David Holm

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]