Re: Best Workstation
On Fri, Oct 29, 2004 at 06:02:36PM -0300, Javier O. Augusto wrote:
> First of all, I know this is _not_ an on-topic question, but I think it
> is way much better fitted than all those crappy spams coming down.
> Maybe one of you is doing the same research as me and can share some
> Well, the fact is that I'm planning to get a new Workstation for one of
> our Labs doing computational fluid dynamics stuff. Got two options, a
> dual Sun Opteron (I could easyly run Debian on it) *W2100z or a dual Sun
> UltraSPARC III (Solaris only) *Blade 2500, same memory (4GiB) and same
> storage (2x73GiB U320).
Check spec.org (http://www.spec.org/cpu2000/results/cfp2000.html).
The Blade 2500 at 1.28GHz gets a similar SPECfp score to an Opteron 140
using software from 2003 (one of AMD's earlier AMD64 submissions).
And I'd get two SATA disks. I know SCSI's better, but it's not _that_ much
better, esp. for a workstation. There's always WD's 10k RPM Raptor drives.
A postdoc here recently got a dual Opteron 246 workstation with 6GB of RAM,
and two SATA disks (set up with software RAID1 and LVM), and it's been doing
very well handling bioinformatics and phylogenetics computations :)
> What would be the best 'bang-of-the-buck' for my labs' necessity?
> Straight horse power.
These days the only question is AMD or Intel (or maybe Mac now that IBM's
CPUs are getting up there in speed). UltraSPARC isn't keeping up:
> PS: Not going after a cluster because of a very tight budget.
A cluster of cheap-o Celerons on mobos with integrated video and lan is
_very_ cheap. It's not scalable, but 10 600$ machines in low-profile cases
sitting in a corner can be useful. (If you've got the admin time and skills
to manage hardware that won't PXE boot, so you need floppy disks and
#define X(x,y) x##y
Peter Cordes ; e-mail: X(peter@cor , des.ca)
"The gods confound the man who first found out how to distinguish the hours!
Confound him, too, who in this place set up a sundial, to cut and hack
my day so wretchedly into small pieces!" -- Plautus, 200 BC