Jan, I agree with your assessment that the need to adjust the memory use per process is a general one in cluster job submission, and that it is in some way implemented by any underlying job management system, and that these extensions ought not to be PBS-specific.
I also looked at your "messy solution". (The code looks very professional, really.) It won't do for my purposes, because I need to present a minimal, easily understood solution. Let me explain my situation: None of the compute resources is under my control. I can point out problems to admins, that is all. I have been assigned two jobs. I and our users are familiar with doing conventional cluster job submission. One job was to bring them into the grid fold, showing them the advantages of globusrun-ws. If it can be shown to be really a cross-platform solution, giving them the ability to (almost) effortlessly switch between grid clusters, the effort will be a success. My other job is to write a report on practical MPI job submission over the grid. We have come a long way, but still have to deal with a couple of practical details. At this point, it looks like both of them will end up as work-arounds to incomplete implementation of a job submission interface in Globus. If with a future release of Globus, these issues can be dealt with, grid job submission will look very attractive to real researchers. Thanks again! On 6.05.08, Jan Ploski wrote: > Steve White <[EMAIL PROTECTED]> schrieb am 05/06/2008 04:43:18 PM: > > > Jan, > > > > It looked good, but unfortunately this is a PBS-only extension, and most > > of our resources (specifically, the one the user in question wants > most), > > are using SGE. > > > > You know, my users say, "Why should I do job submission via Globus? > > Why not just gsissh in and do a job submission to the cluster's job > > manager?" > > > > My answer has been "Ah, because the Globus way abstracts away all the > > details of the cluster's job manager---you just write one job > description > > for Globus, and voilà, it runs on all Grid clusters". > > That's the main idea, yes. > > > If in fact the user has to write different job descriptions for each job > > manager on each cluster... > > That's the sad reality. > > > ... what is the correct answer to this question? > > When I recommended it, I didn't pay attention to the fact that this was > PBS-only. In fact I see nothing in the elements' meaning which would make > it impossible to implement in other LRMS adapters (I don't know SGE, but I > know Condor). Yet, strangely, the extension is called "PBS Node Selection > Parameters" in the documentation. Perhaps someone from Globus can comment > whether this extension is going to be generalized? > > If you are dealing with a small number of resources under your control, it > might be reasonable to just hack the SGE adapter (Perl) to support this > extension and possibly even contribute it back to Globus... That way you > could keep your promise to the users. > > Regards, > Jan Ploski > > -- > Dipl.-Inform. (FH) Jan Ploski > OFFIS > Betriebliches Informationsmanagement > Escherweg 2 - 26121 Oldenburg - Germany > Fon: +49 441 9722 - 184 Fax: +49 441 9722 - 202 > E-Mail: [EMAIL PROTECTED] - URL: http://www.offis.de > - - - - - - - - - - - - - - - - - - - - - - - - - Steve White +49(331)7499-202 e-Science / AstroGrid-D Zi. 35 Bg. 20 - - - - - - - - - - - - - - - - - - - - - - - - - Astrophysikalisches Institut Potsdam (AIP) An der Sternwarte 16, D-14482 Potsdam Vorstand: Prof. Dr. Matthias Steinmetz, Peter A. Stolz Stiftung privaten Rechts, Stiftungsverzeichnis Brandenburg: III/7-71-026 - - - - - - - - - - - - - - - - - - - - - - - - -
