Date: Thu, 6 Nov 2003 20:43:30 -0800 From: Brooks Davis <brooks@one-eyed-alien.net> To: Craig Rodrigues <rodrigc@crodrigues.org> Cc: Rayson Ho <raysonlogin@yahoo.com> Subject: Re: Linux vs FreeBSD clusters (was: how are the Redhat product changes affecting existing and future plans?) Message-ID: <20031107044322.GA23181@Odin.AC.HMC.Edu> In-Reply-To: <20031107040415.GA5711@crodrigues.org> References: <20031106145623.GA5867@iib.unsam.edu.ar> <20031107005951.2157.qmail@web11407.mail.yahoo.com> <20031107040415.GA5711@crodrigues.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--a8Wt8u1KmwUX3Y2C Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Nov 06, 2003 at 11:04:15PM -0500, Craig Rodrigues wrote: > Not quite the same as an HPC cluster, but take > a look at the University of Utah's Emulab: >=20 > http://www.emulab.net >=20 > It is heavily based on FreeBSD (i.e. makes use of FreeBSD routing, > Dummynet, etc.). The Emulab is a remotely accessible testbed > that researchers can use to conduct network experiments. It > consists of about 200 PC nodes. The same company that=20 > Brooks works for (Aerospace), has apparently set up > an internal testbed based on the Emulab software developed at Utah. We (my department, but mostly different people then Fellowship) have a small 10-node setup (though each node does have 6 gigabit ports :-). I think we're aiming to upgrade to around 48 nodes in the next year. Our HPC cluster is currently pretty close to what's described in the paper, though we are up to 160 nodes and we're adding rack space for another 192 this year. The short version of my take on which OS to run on your cluster is that so long as it runs the apps you need, the best OS is one you know how to admin well since that's most of the work. I've spent a few weeks here and there porting applictions or improving their ports, but by and large, most key systems are already ported to the major UNIX platforms. The free MPI implemntations work on just about anything, the base Ganglia metrics work nearly everywhere (FreeBSD and Linux are at feature parity in the upcoming release), and SGE works on a wide range of platforms. On an amusing note, we were the launch customer for Grid Mathematica despite not running a supported OS because the Linux version runs just fine on FreeBSD. -- Brooks --=20 Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4 --a8Wt8u1KmwUX3Y2C Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQE/qyLlXY6L6fI4GtQRAgRdAKDL9Eh1RB4Q0eBBFhKNhizmT9+cbQCeIe0+ VNFsAkXmJmeO9w7ESxHEuvE= =yKvI -----END PGP SIGNATURE----- --a8Wt8u1KmwUX3Y2C--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20031107044322.GA23181>