VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Wed Aug 27 2003 - 13:39:17 CDT

Hi guys,

On Wed, Aug 27, 2003 at 02:24:11PM -0400, Fred Salsbury wrote:
>
> Hello --
>
> Thanks for your reply. I would be using this for visualization and some
> data analysis; not for production runs, so it is not clear how useful your
> benchmarks are for this purpose -- though I bookmarked them for future
> reference on production machines.
>
> John recommended as SGI/SUN workstation.

Let me state that this is versus spending money on Xeon boxes, since their
price point is in the same general area as the workstations. You could get
a good performing Athlon PC for much less. If money is no object, then yeah
I'd definitely get a real workstation rather than buy a PC, particularly
if you care about things like stereoscopic display, etc.

> I actually have mostly used VMD on O2's.

My own experience with the O2 and Indy machines was not very positive,
but I've found that VMD ran well on Octane/Octane2, Onyx, and Indigo2
machines in the past. I suspect that their "Fuel" workstation is pretty
decent, but I have not tested one personally.

> Does anyone have experience with the SGI Fuel workstation vs
> a dual processor Xenon workstation or a dual processor Athalon
> workstation, or even a dual Operton workstation?

Of those 3 (Fuel, Xeon, Opteron), the Opteron would be most interesting
to me, but I think that NVidia is the only vendor that currently has drivers
that run on the Opteron platform, so that may curtail that choice somewhat.

At Siggraph 2003 I ran some Tachyon performance tests on a cluster of dual
processor Opterons running at 1400MHz and they annihilated all of the other
PC type systems I've previously benchmarked, I was VERY impressed.
I attribute the great performance I got to the memory system on the
Opterons I tested. Tachyon is a very memory-bound application, its
sensitive to both latency and bandwidth, and the Opteron performed
very well in all of the tests I ran. It was also very easy to compile
Tachyon in 64-bit mode on that platform. The machines I tested were
running Linux with one of the bleeding edge versions of gcc.

  John 

-- 
NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
Email: johns_at_ks.uiuc.edu                 Phone: 217-244-3349              
  WWW: http://www.ks.uiuc.edu/~johns/      Fax: 217-244-6078