charm++ how to handle SMP/Multicore

From: Bjoern Olausson (namdlist_at_googlemail.com)
Date: Sun Aug 09 2009 - 17:01:48 CDT

Hi all,

we have a Cluster build on new AMD Hexacores (Six-Core AMD Opteron 2427) with
two Hexacores per Node. The cluster should mainly run NAMD

The 24 Nodes are connected to each other with DDR Infiniband.

So now my question:
Which Option should I choose when
        1) Compiling charm with mpi (mvapich2) to run on >12Cores
        2) Compiling a version that should perform best on one node <=12Cores

These are the Options Charm provides:
         1) single-threaded [default]
         2) multicore(single node only)
         3) SMP
         4) POSIX Shared Memory

Thanks a lot for any suggestions.

Kind regards
Bjoern

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:51:20 CST