[Wien] lapw0 in parallel

[Jorissen Kevin]

I do not think 15 minutes is unreasonable for lapw0 of a large case.
 
As to the remark I posted earlier, I think I formulated it rather unclearly.  I have sometimes experienced that many lapw0's running on a server somewhat become a bottleneck, and then it is useful to just run them remotely on a node and keep the server free.  This is not parallellization, it is merely working remotely.  Of course, contrarily to the MPI case, this will not intrinsically make lapw0 faster (perhaps even slower, if your nfs really sucks).  It will only use your hardware more efficiently.
That is the only point I was trying to make (and, indeed, the only thing the mentioned script will do to help you).
 
 
Kevin Jorissen
 
EMAT - Electron Microscopy for Materials Science   (http://webhost.ua.ac.be/emat/)
Dept. of Physics
 
UA - Universiteit Antwerpen
Groenenborgerlaan 171
B-2020 Antwerpen
Belgium
 
tel  +32 3 2653249
fax + 32 3 2653257
e-mail kevin.jorissen at ua.ac.be
 

________________________________

Van: wien-admin at zeus.theochem.tuwien.ac.at namens Michael Gurnett
Verzonden: vr 22-10-2004 18:18
Aan: wien at zeus.theochem.tuwien.ac.at
Onderwerp: Re: [Wien] lapw0 in parallel



Its a big system (70 atoms). I'm running the compiler options given for the
I option in siteconfig using mkl 6.1

-----Original Message-----
From: "L. D. Marks" <L-marks at northwestern.edu>
To: wien at zeus.theochem.tuwien.ac.at
Date: Fri, 22 Oct 2004 09:05:44 -0500 (CDT)
Subject: Re: [Wien] lapw0 in parallel

> 15 minutes is really rather long. Are you compiling lapw0 with
> optimizations turned on? For some of the routines this can have a big
> effect, and I think lapw0 may be one of them.
>
> On Fri, 22 Oct 2004, Stefaan Cottenier wrote:
>
> >
> > >I'm running a small linux cluster (7 computers) which is working
> > flawlessly. However, 1/3 of
> > >the time is taken in lapw0. I was therefore wondering if I should
> start
> > running lapw0 in
> > >parallel. Must I install MPI to do this, and is it worth it (lapw0
> take 15
> > minutes at the
> > >moment). Would be nice if someone can give me instructions on how to
> > proceed.
> >
> > Yes, you need MPI for that. But the easiest work-around is: if you
> have
> > several independent jobs of this type, then start several of them
> > simultaneously, not on 7 but on less machines. That will decrease the
> > effective idle time of each machine. Ideally, if you can start 7
> serial
> > jobs, there will be no idle time at all.
> >
> > Stefaan
> >
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >
>
> -----------------------------------------------
> Laurence Marks
> Department of Materials Science and Engineering
> MSE Rm 2036 Cook Hall
> 2225 N Campus Drive
> Northwestern University
> Evanston, IL 60201, USA
> Tel: (847) 491-3996 Fax: (847) 491-7820
> mailto:L-marks at northwestern.edu
> http://www.numis.northwestern.edu
> -----------------------------------------------
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien


_______________________________________________
Wien mailing list
Wien at zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/ms-tnef
Size: 7129 bytes
Desc: not available
Url : http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20041022/d4bc5c18/attachment.bin