Subject: Re: clustering
To: None <port-i386@netbsd.org>
From: gabriel rosenkoetter <gr@eclipsed.net>
List: port-i386
Date: 04/16/2002 14:21:15
--fmvA4kSBHQVZhkR6
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

[Adam, please wrap your lines at a sane margin, where "sane" is
typically defined by Unix folks as < 80 columns. Thanks.]

On Tue, Apr 16, 2002 at 10:10:51AM -0700, Adam Russell wrote:
> I have an application which has matured to the point where I feel
> that it would be advantageous to run over multiple CPUs. Could
> someone point me to the veritable cache of information on clustering
> with NetBSD which must surely exist but somehow has eluded my
> fevered searching?

Have a look in pkgsrc/parallel.

clusterit works pretty well for batch-style jobs, glunix may not
work at all right now (last I heard, Berkeley ditched the dist files
and I have yet to dig up local copies to put on ftp.netbsd.org, if I
even still have local copies), mpi-ch and pvm3 both require that you
learn how to use their parallel library stuff, but both work quite
well for either batch or communicating parallel processes.

(This is all pretty standard Unix clustering stuff that's been
around far longer than Beowulf and is far more general. If you want
to learn more, go read the myriad academic papers on MPI, PVM, and
GLUnix. clusterit is a bit of an exception, it's loosely based on
GLUnix but oriented towards just getting some batch processing done,
and was written by the maintainer of the GLUnix package.)

--=20
gabriel rosenkoetter
gr@eclipsed.net

--fmvA4kSBHQVZhkR6
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (NetBSD)
Comment: For info see http://www.gnupg.org

iEYEARECAAYFAjy8a5sACgkQ9ehacAz5CRpEgwCeONCygaBnFwhzuPnzZY1wNNo3
gB4AnjTuon/FYeIOC5Bf1KsK3YQ33iLF
=CdpX
-----END PGP SIGNATURE-----

--fmvA4kSBHQVZhkR6--