[mpich-discuss] Scalability of Intel quad core (Harpertown) cluster

[Pavan Balaji]

Is the application itself scalable? It is possible that the application 
behavior itself might have some inherent scalability issue and might 
have nothing to do with communication. You should consider profiling 
your application to see what exactly is happening before trying out 
different communication channels, MPI implementations or networks.

MPE/Jumpshot (that is packaged with MPICH2) is a good tool to profile 
the application code. mpiP is another tool to see how much of the 
application time is actually spent inside MPI. For example, if the 
application is only spending 5% of its time within MPI, irrespective of 
what network or communication channel you use, it'll not help you much.

  -- Pavan

On 03/28/2008 12:50 AM, Hee Il Kim wrote:
> Hi,
> 
> I recently made a Rocks cluster of 10 comuting nodes with dual Intel 
> quad core cpus (Harpertown). The nodes are connected with gigabit networks.
> 
> The problem is I got bad scalability from the benchmark test of Cactus 
> which is well-known program for numerical relativity. It might be 
> natural with the gigabit network. However, I suspect I missed something 
> important or it's a generic problem of clusters of multicore cpus 
> because the benchmark problems does not require massive communication 
> between the computing nodes.
> 
> I tested openmpi, mpich as well as mpich2 but no significant differences 
> between openmpi and mpich2. The mpich2 was configured with ssm. After 
> reading some documents on nemesis and multicore cpus, I tested with 
> nemesis but got worse results than with ssm.
> 
> - Is there any optimized configuration option for multicore cpus like 
> Harpertown?
> - Could it be improved only with infiniband, myrinet,...?
> - If the gigabit network was the cause, could it be improved with Open-MX?
> 
> I'm a newbie on this field. The questions must be not clear to you. I 
> appreciate any helps in advance.
> 
> Kim, Hee Il

-- 
Pavan Balaji
http://www.mcs.anl.gov/~balaji