[Beowulf] Problems scaling performance to more than one node, GbE
a28427 at ua.pt
Mon Feb 16 07:43:35 PST 2009
On Sat, Feb 14, 2009 at 10:13 AM, Nicholas M Glykos <glykos at mbg.duth.gr>wrote:
> Hi Tiago,
> > I tried with Gromacs ...
> Concerning your MD tests, would it be worthwhile to check also NAMD's
> parallel efficiency ? Based on past experience, I would suggest that you
> try both the UDP- & TCP-based versions (don't forget to use the +giga and
> possibly the +atm flags). Also, keep in mind that parallel efficiency
> varies greatly with problem size (for example, you should expect very
> different results from, say, a 1,500 atom peptide system, as compared with
> a 100,000 atoms large protein system).
I will try NAMD like you say. Can you give me a usual performance test you
have, to which I can compare to? Gromacs tests I've been able to scale
better are the ones with 100k+ atoms.
> > From what I could also digg around, it seems that some switches have too
> > much latency and hamper any kind of proper performance from GbE. Were
> > my case, any benchmarks I could use to test that theory out?
> We have been using NetPipe (http://www.scl.ameslab.gov/netpipe/) to
> measure latency, but the experts on this list will probably tell you more
> about the subjest that you'd ever wished to know ;-)
> Ok, thanks! Much appreciated.
> Dr Nicholas M. Glykos, Department of Molecular
> Biology and Genetics, Democritus University of Thrace,
> University Campus, 68100 Alexandroupolis, Greece, Fax +302551030620
> Tel ++302551030620 (77620), http://www.mbg.duth.gr/~glykos/
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Beowulf