Re: AW: AW: AW: NAMD crashes with OpenMPI/OpenMX

From: Jim Phillips (jim_at_ks.uiuc.edu)
Date: Wed Jun 15 2011 - 11:17:47 CDT

I actually consider apoa1 a good size for scaling tests because it shows
scaling differences between machines or whatever settings you're trying to
tune. He only has 18 cores, so the problem isn't system size.

-Jim

On Wed, 15 Jun 2011, Norman Geist wrote:

> Oh, much better ping than mine. Oh, and Apoa is a mini system and isn't very
> good for scaling testing. Try the stmv ;)
>
> Let me know.
>
> Norman Geist.
>
>
> -----Ursprüngliche Nachricht-----
> Von: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] Im Auftrag
> von Thomas Albers
> Gesendet: Mittwoch, 15. Juni 2011 17:01
> An: namd-l_at_ks.uiuc.edu
> Betreff: Re: AW: AW: namd-l: NAMD crashes with OpenMPI/OpenMX
>
> On 6/15/2011 10:37 AM, Norman Geist wrote:
>> Well I have a TCP version, and yes tcp congestion control highspeed made a
>> big deference in scaling behavior on my cluster. Maybe try out. What is
> the
>> ping time of your cluster?
>
> While running the ApoA1 benchmark on 18 cores, 4 computers:
> --- ferrari ping statistics ---
> 51 packets transmitted, 51 received, 0% packet loss, time 50011ms
> rtt min/avg/max/mdev = 0.035/0.489/4.324/0.918 ms
>
> no load:
> --- ferrari ping statistics ---
> 11 packets transmitted, 11 received, 0% packet loss, time 9999ms
> rtt min/avg/max/mdev = 0.052/0.052/0.054/0.007 ms
>
> Thomas
>
>

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:20:27 CST