Edward d'Auvergne wrote:
On 5/29/07, gary thompson <garyt.and.sarahb@xxxxxxxxx> wrote:
Hi Ed
On 5/29/07, Edward d'Auvergne < edward.dauvergne@xxxxxxxxx> wrote:
> On 5/18/07, Gary S. Thompson <garyt@xxxxxxxxxxxxxxx> wrote:
> > Edward d'Auvergne wrote:
> >
> > >On Fri, 2007-05-04 at 13:59 +0100, Gary S. Thompson wrote:
>
> [snip]
>
> > >If using '-np 6', shouldn't the number of slaves be 6?
> > >
> > nope there needs to be one processor which is the master and you
just
> > tell mpi how many prcoessors you want ( I will investigate
running jobs
> > on the master in a thread at some point (maybe never depending
;-)) but
> > this places extra requirements on the mpi implimentation and is
thus a
> > special case, I can give more details if you want me to)
>
> I would personally avoid running one of the calculations on the
master.
In general yes I agree, however, if you have a small cluster this is
quite a
waste
Can this be done as a slave MPI process on the master?
Edward
.
not easily ;-( unfortunatley and the methodolgy you use will be
implimentation dependant. Take for example LAM MPI this uses a file of
machines to use and you can enter the same machine twice (thi is how
you test on a single processor machine and deal with dual processor
machines) However there is no guaruntee that machine which has rank 0
(which I always use for the maste) will on the machine that we have
selected to run two prcoesses on....
I think the way round this is threads
regards
gary
--
-------------------------------------------------------------------
Dr Gary Thompson
Astbury Centre for Structural Molecular Biology,
University of Leeds, Astbury Building,
Leeds, LS2 9JT, West-Yorkshire, UK Tel. +44-113-3433024
email: garyt@xxxxxxxxxxxxxxx Fax +44-113-2331407
-------------------------------------------------------------------