Re: openMosix + MPI + PBS

Talimades <[email protected]>
Newsgroups gmane.linux.cluster.openmosix.general
Message-ID <[email protected]>
Dirk Eddelbuettel wrote:

>On 18 December 2005 at 12:35, Talimades wrote:
>| Hello,
>| I have a cluster with openMosix (on Suse 9.0).
>| As I have understood openMosix takes charge of the efficient 
>| distribution of the work.
>| But on the other hand when MPI wants to be used it is necessary to say 
>| those who are the nodes that will be used.
>
>I believe this is covered in the FAQ or openMosix wiki: You can combine the
>two: tell MPI to launch 'N' instances. If you do that in an openMosix cluster
>with 'M' computers, you are likely to end up with N/M instances per node (for
>as long as the N instances are roughly equal, and your computing nodes are
>identical; otherwise openMosix adjusts).
>
>I have a few slides from talks at Usenix 2004 and DSC 2005 that talk about
>this. I often use a higher level wrapper for R called SNOW (for 'Simple
>Network Of Workstations') which can uses MPI, PVM or plain old sockets.
>
>| I would like to know which is the best way of doing that the users put 
>| works in queue.
>| I know that exists openPBS, but I really don't know how it works, what 
>| is, what allows to do and what not, etc.
>| 
>| Can someone orientate me?
>| Some good link to learn of all that?
>
>You may benefit from reading a book like Sloan's 'High Performance Linux
>Clusters with Oscar, Rocks, openMosix and MPI'. 
>
>Hope this helps, Dirk
>
>  
>
Thank you.


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.