[Bioclusters] opteron vs xeon quad cpu boxes

[Joe Landman]

On Tue, 6 Jul 2004 15:35:15 +0000 (UTC), Ognen Duzlevski wrote
> Hi,
> 
> we are looking at a quad-cpu server for our bioinformatics apps like 
> multithreaded clustalw, blast and some in-house multithreaded 
> application I developed over time.
> 
> I noticed that I can get a basic 8GB RAM 4xOpteron 848 with a RAID 
> controller and a couple of hdds for close to $17,000. At the same time 
> it looks like I can get a 4xXeon MP 2.7Ghz with the same ram/hdd 
> configuration for close to the same price.

Really (quad opteron price)?  Parts alone for those units in that config will be
 pretty close to $16000.  We have built a few custom units for various
customers.  The problem is memory pricing.  Keeps fluctating.  Check the actual
config pricing with the vendor and get a quote. 

> 
> Any insights on the performance of the Opteron quad boxes? How do they 
> compare to their 32-bit counterparts like Xeon MPs? Have they proven 
> suitable for your bioinformatics environment?

In general, yes, the Opterons should be better than the quad Xeon (though I have
not run on a quad Xeon, so I cannot really compare).  This is due to the quad
Xeon's sharing a memory bus, and thus having a highly limited bandwidth to RAM,
and the Opterons getting one 128 bit path and independent memory controller per
CPU.  

CPU to CPU, the Opterons are somewhat faster (running in 64 bit mode) even with
a 1 GHz clock difference on BLAST, HMMer, and a few others.  The intel should be
faster on some codes which depend only upon the core CPU instruction issue rate
and do not touch RAM all that much.  There are few of those, though they tend to
be Monte Carlo type codes.

> 
> Any advice would be valuable!

We would be happy to talk to you about this (offline so as not to spam everyone).

Joe

> 
> Thanks,
> Ognen
> 
> --
> Ognen Duzlevski
> Digital Biology Laboratory
> 302 North Engineering Building
> University of Missouri-Columbia
> (573) 882-5978
> _______________________________________________
> Bioclusters maillist  -  Bioclusters@bioinformatics.org
> https://bioinformatics.org/mailman/listinfo/bioclusters


--
Joseph Landman, Ph.D
Scalable Informatics LLC,
email: landman@scalableinformatics.com
web  : http://scalableinformatics.com
phone: +1 734 612 4615