Charm++ parallel programming system

[Pritish Jetley <pjetley2 AT illinois.edu>]

Hello Anirban,

That particular test program is an adaptation of the 'barnes'
benchmark in the SPLASH suite of programs. It is meant to work only on
single-node, shared-memory machines. In order to get it to work, you
must make a 'multicore-linux64' build of Charm++. If you are looking
for a distributed memory implementation of the Barnes-Hut algorithm,
you might want to check out ChaNGa
(http://software.astro.washington.edu/nchilada/).

Pritish

On Tue, Jan 25, 2011 at 10:59 AM, Anirban Chatterjee
<achatter AT cse.psu.edu>
 wrote:
> Hi,
>
> I compiled charm++ and barnes-charm on a linux cluster. The compilation
> seems to work without an error.
> However, when I try running it on 4 nodes (1 core on each node) then
> only process 0 seems to have the "bodytab" (see the functions
> Main::startSimulation in Main.cpp and ParticleChunk::SlaveStart in
> ParticleChunk.cpp).
> I get a Signal 11 on all other processes other than Process 0 when they
> try to access bodytab.
>
> Also this program runs fine on 1 processor.
>
> Did anyone see this error earlier? I don't know what could be going wrong.
>
> Thanks,
> Anirban
>
> --
> Anirban Chatterjee
> 350 IST Building,
> Department of Computer Science&  Engineering,
> The Pennsylvania State University,
> University Park, PA 16802
>
> _______________________________________________
> charm mailing list
> charm AT cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/charm
> _______________________________________________
> ppl mailing list
> ppl AT cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/ppl
>



-- 
Pritish Jetley
Doctoral Candidate, Computer Science
University of Illinois at Urbana-Champaign