[Beowulf] Keeping the Athlon MP cluster limping along

[Robert G. Brown]

On Thu, 9 Dec 2004, Bari Ari wrote:

> Robert G. Brown wrote:
> 
> > We have a stack (or had, we may finally have pitched them) of close to
> > 100 of these pieces of crap.  These are what we bought the 2466's to
> > replace, as they are actually BETTER than 2460's.  They still haven't
> > got the bios right in the 2466's, but the 2460 bios -- just don't get me
> > started.
> 
> Have you looked at using LinuxBIOS for them? Tyan has a full-time 
> LinuxBIOS developer. He posts near daily on the 
> Linuxbios at clustermatic.org list and I hear that LANL uses plenty of them.
> 
> Bari Ari

We just want all the 246x systems to go away.  When the hardware doesn't
break, we've got the cluster set up so that it will boot and run
reasonably well (the systems are, or were, very admirable performers
numerically).  At this point, it is human interaction and human time
wasted on hardware problems that are the real expense.  If it weren't
opportunity cost time vs money we don't have, we would have thrown the
entire cluster away long ago and replaced it.

    rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu