[Dock-fans] Question re MPI Dock

Philip Crawford p.crawford at unsw.edu.au
Wed May 17 23:19:18 PDT 2006

Hi. I have come up with what I'm hoping is a workaround but will not test
until the weekend. Doing this stuff in my own time.
What I'm going to try is to give the full path to the working directory for
picking up the .in file and also for writing, as in:
mpirun -np xx dock5mpi -i $PBS_O_WORKDIR/dock.in -o $PBS_O_WORKDIR/dock.out
I also intend to have the dock.in file also reference the parameter files by
full path.
It seems that these files are written by the active processes. I think that
the master process needs the database and the ranked and scored output files
in the local scratch area.
I'll try to remember to let you know how this goes. 
Philip Crawford, B. Comp. Sc., MIEEE
School of Medical Sciences
The University of NSW
Phone: +61-2-9385 2564
Mobile: +61-419-294 698
Fax: +61-2-9385 1059
Email: p.crawford at unsw.edu.au


From: Kimberly Begley [mailto:k.begley at griffith.edu.au] 
Sent: Thursday, 18 May 2006 2:36 PM
To: Philip Crawford
Subject: Re: [Dock-fans] Question re MPI Dock

Hi Philip, 
I'm working with DOCK on our cluster here at Griffith Univeristy in Brisbane
- just wondering if you got it working over there - our software devleoper
wrote a shell scrip to copy the input file to the nodes - you can also put
it in your PBS script - so yes each node needs to be able to see it. We got
that working but are having some segmentation faults and I'll be curious if
this occurs on your system as well. I'll pass on the email address of the
developer - he's becoming an expert with DOCK 5.3 /5.4 software
installations - His name is Othmar Korn and email is o.korn at griffith.edu.au.

Hope that helps 

Kimberly Begley
Research Fellow, Bioinformatics

Research Computing Services
Gold Coast Campus, Griffith University
58 Parklands Drive, Southport
Queensland 4215, Australia
k.begley at griffith.edu.au

"Philip Crawford" <p.crawford at unsw.edu.au> 
Sent by: dock-fans-bounces at docking.org 

18/05/06 10:02 AM 

<dock-fans at docking.org> 


[Dock-fans] Question re MPI Dock


Hi. Having a few problems with MPI Dock on large cluster. 
The administrators request that jobs be run from a local folder, set up by
PBS, on the local file system of the Master node. This is to avoid slowdowns
on the head node. As far as I can tell, this local scratch space is not
accessible via NFS from the other execute nodes. 
It seems to me that Dock 5 MPI version actually has each node read the
dock.in file and produce its own separate dock.out.nodeno file, though the
manual says that "The Master node performs file processing and input/output,
whereas the slaves perform the actual calculations". 
In my case, when I run from this local scratch space, the execute directory
is not accessible from the other execute nodes and only the local processes,
in my case dock.out and dock.out.16 had been written. Though I had been
allocated 15 other DP nodes, Dock was not running on those I logged onto. 
I'd like to confirm that I need to run the MPI job from a working directory
which is accessible from the other execute nodes. Am I correct? 
Thanks in advance, 
Phil Crawford 
Philip Crawford, B. Comp. Sc., MIEEE 
School of Medical Sciences 
The University of NSW 
Phone: +61-2-9385 2564 
Mobile: +61-419-294 698 
Fax: +61-2-9385 1059 
Email:  <mailto:p.crawford at unsw.edu.au> p.crawford at unsw.edu.au 
Dock-fans mailing list
Dock-fans at docking.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://blur.compbio.ucsf.edu/pipermail/dock-fans/attachments/20060518/ce205991/attachment.html

More information about the Dock-fans mailing list