Re: [AMBER] how many cores should be used in MMPBSA.py.MPI calculations at most

From: Qinghua Liao <fantasticqhl.gmail.com>
Date: Mon, 12 Dec 2011 13:42:06 +0100

Dear Bill,

Thanks very much for your explanations. It is helpful. Thanks.

Best Regards,

Qinghua

On Mon, Dec 12, 2011 at 1:36 PM, Bill Miller III <brmilleriii.gmail.com>wrote:

> In general, the more cores the faster the calculation gets done. There is
> no significant decrease in efficiency by using more cores. MMPBSA.py uses
> serial sander for each processor and splits the number of frames as evenly
> as possible amongst the processors. However, if you are performing normal
> mode analysis be mindful of the available memory because using too many
> processors sharing the same memory can result in insufficient resources and
> the job dying.
>
> I hope that helps.
>
> -Bill
>
>
>
> On Dec 12, 2011, at 4:31 AM, Qinghua Liao <fantasticqhl.gmail.com> wrote:
>
> > Dear amber users,
> >
> > When doing mm/pbsa calculations using the python version, usually how
> many
> > cores should I use at most while it can be still efficient
> parallelisation?
> > Thanks very much!
> >
> > Best Regards,
> >
> > Qinghua
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Dec 12 2011 - 05:00:03 PST
Custom Search