Re: Namd benchmark problem

From: Joshua Adelman (jadelman_at_berkeley.edu)
Date: Sun Oct 25 2009 - 08:18:23 CDT

If you look at the config file that you ran, you'll see that you ran
500 steps at 1 fs/step, so you simulated half a picosecond. A quick
back of the envelope calculation says that if you took about 4 minutes
to run 500 steps, then it would take you about 5.6 days to run 1
nanosecond. But you don't need to resort to doing this by hand . . .

If you look at is the output in the log file that approximates
information like the number of nanoseconds/day. As Axel stated, the
whole point of a benchmark is to be able to run the same benchmark
many times with different numbers of processors in a short period of
time, so that you can predict what sort of scaling you might achieve
on your system.

Also if you take a careful look at the webpage you refer to, you will
find a paragraph with information that you should read, and would
probably clear up a lot of your misunderstanding.

Best wishes,
Josh

On Oct 25, 2009, at 2:46 AM, Sangamesh B wrote:

> Dear Sir,
>
> On Sat, Oct 24, 2009 at 10:00 PM, Axel Kohlmeyer
> <akohlmey_at_gmail.com> wrote:
> On Sat, 2009-10-24 at 12:42 +0530, Sangamesh B wrote:
> >
> > Dear Namd users,
> >
> > I've installed NAMD on a Linux HPC cluster by both ways (1)
> > Downloading binaries (2) By compiling the source with Open MPI.
> >
> > To test the scaling of Namd, I downloaded the standard benchmark
> apoa1
> > from NAMD website. After doing the benchmark I found that its taking
> > very little time 3-4 minutes to complete(with 4 cores), but on the
> > website its shown its taking 4days to complete.
>
> please explain where it says that it takes this long?
>
> The graph shown on the following webpage
>
> http://www.ks.uiuc.edu/Research/namd/performance.html
>
> From this graph I understood - 1 week for 4 processor job, 4 days
> for 8 processor job etc
> the input does just a few hundred steps. any benchmark
> that would need 4 days to complete is very impractical.
>
> Also the graph says, namd scales upto 256 cores. But in may case,
> its taking very less time and not scaling beyond 24 cores.
> > May I know what mistake am doing here?
>
> you are must be misreading something.
>
> > On the website, there is no info provided about how to do the
> > benchmark.
>
> you just run it like any other namd input.
>
> for a bigger problem size namd input, upto what number of cores will
> NAMD scale?
> axel.
>
> > Thanks
> >
> >
> --
> Dr. Axel Kohlmeyer akohlmey_at_gmail.com
> Institute for Computational Molecular Science
> College of Science and Technology
> Temple University, Philadelphia PA, USA.
>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:51:36 CST