From: L. Michel Espinoza-Fonseca (mef_at_ddt.biochem.umn.edu)
Date: Mon Feb 12 2007 - 11:19:57 CST
Hi Peter,
Thank you for your reply. Certainly, it'll be awesome to have the
improved version of the program. I was also kind of surprised that it
is taking such a long time to calculate the grids, specially on such
powerful machines.
And yes, the Blade is a shared memory machine (or I think so :)).
Cheers,
Michel
2007/2/12, Peter Freddolino <petefred_at_ks.uiuc.edu>:
> Hi Michel,
> the rate limiting step for ionize is the initial grid calculation, not
> the ion placements. Once your grid is done the ion placements are
> practically free. As of VMD 1.8.6 cionize is going to be included with
> VMD (similar to catdcd); in the meantime I'll send you a newer version
> off list that should give you somewhat better performance and give you
> more detailed timing and processor usage info so you can make sure it's
> using all the processors. My guess would be that a system that big
> should have finished in 2 days on 24 processors, which is why I'm puzzled...
> BTW, if the bladecenter a shared memory machine? If not you certainly
> won't be running in parallel.
> Peter
>
> L. Michel Espinoza-Fonseca wrote:
> > Hi,
> >
> > I've started to use the IONIZE program (kindly provided by Peter
> > Freddolino) to add ions to a 2-million atom system. The net charge of
> > my system is -3,600 e. I tried to neutralize such system by using 50
> > processors (Altix) for 2 days, and nothing happened. Then I tried to
> > do it on 100 processors, and the job was killed. Now I'm trying to do
> > it with 500 processors (BladeCenter, AMD Opteron) and I'm waiting to
> > see if this time the job will be completed. So here is my question: do
> > you have any idea of how much time it could take to add such a big
> > amount of ions?
> >
> > Thanks!
> >
> > Michel
>
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:44:23 CST