[gmx-users] BENCHMARK RESULTS

[Erik Lindahl]

Hi,


My guess would be that Mosix is causing the problems - that approach is 
really orthogonal to MPI parallelization....

I know that a couple of people on the list use it though, so check if 
any of them see good scaling. (You might need to write a mail with that 
subject though, since the traffic is pretty high :-)

Cheers,

Erik

On Tuesday, October 21, 2003, at 04:27 AM, Andrea Carotti wrote:

> Hi all,
>     I' ve tested my system with the dppc test using the -shuffle  
> -sort flags.
> My system is composed by 5 cpus ATHLON XP 2400+ running oMosix on RH8, 
> NIC 10/100. I've obtained this results:
>                     (Mnbf/s)    (GFlops)     (ps/NODE hour)     (NODE 
> hour/ns)
>
> Performance: 42.782     1.543                13.211                    
> 75.694
>
> Load imbalance reduced performance to 500% of max
>
> Total Scaling: 98% of max performance
>
> I'll really appreciate an opinion and eventually also some 
> hints on how I can optimize my system.
> Thanks in advance
> Andrea
>  
>
> #############
> Andrea Carotti
> Dip. Farmaco-chimico
> Università di Bari
> via E. Orabona, 4
> CAP 70124
> e-mail : and.carotti at farmchim.uniba.it
> web: http://www.farmchim.uniba.it/samba/
> phone: +39080/5442638
> fax: +39080/5442230
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 1906 bytes
Desc: not available
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20031021/700079a0/attachment.bin>