Re: subarchitectures
Greetings, and thanks for your reply!
Erik <journey@jps.net> writes:
> --dDRMvlgZJXvWKvBx
> Content-Type: text/plain; charset=us-ascii
> Content-Disposition: inline
> Content-Transfer-Encoding: quoted-printable
>
> On Fri, Dec 01, 2000 at 03:06:42PM -0500, Camm Maguire wrote:
> >=20
> > Greetings! The good news is that in days, we should have some
> > optimized atlas blas libraries for a variety of recent processors.
> > The challenge is how to adapt the package -- when built on an Athlon,
> > PIII, and PII, the resulting binaries will run only on those
> > subarchitectures, (and much faster!) I suppose all binaries could be
> > included in the package, and only the appropriate one installed, but
> > how does this jive with the automatic build daemons? Such a package
> > would have to be partially built on several different machines. On
> > the other hand, I could have several different packages, each of which
> > would build only on a given subarch, and would produce a distinct
> > package name. That might not work so well with automated build
> > daemons either. =20
> It doesn't actualy need to be built on different machines. What happens
> is your configure script auto-detects the processor and optomizes for it.
> You just need to specificaly tell it what processor to build for and it
> will ignore the current processor.
> >=20
The problem here is that the standard atlas compilation procedure both
compiles *and executes/times* the resulting code on the building
machine. It thereby selects the fastest routines for that box. To my
knowledge, there is as yet no 'cross-compilation' capability in atlas,
although I'm clarifying this with the author. Under the assumption
that the compiled code must be run and timed on the compilation
machine, what is the best way to handle this situation?
Thanks!
--
Camm Maguire camm@enhanced.com
==========================================================================
"The earth is but one country, and mankind its citizens." -- Baha'u'llah
Reply to: