Minutes of today telecons + reminder tomorrow telecon 3 PM GMT
Dear all, in the following page you can find links to the minutes of the two today telecons: http://forge.ogf.org/sf/wiki/do/viewPage/projects.glue-wg/wiki/PhoneMeeting2... the discussion on Storage entities will be continued Fri 25th 3 PM GMT. The next telecon is for tomorrow and we'll resume discussion on policy and values when unknown. http://forge.ogf.org/sf/wiki/do/viewPage/projects.glue-wg/wiki/PhoneMeeting2... Cheers, Sergio -- Sergio Andreozzi INFN-CNAF, Tel: +39 051 609 2860 Viale Berti Pichat, 6/2 Fax: +39 051 609 2746 40126 Bologna (Italy) Web: http://www.cnaf.infn.it/~andreozzi
Hi Sergio, Here are the use cases you asked me to send the list. Let me know if you require us to have a telecon to talk about the use cases. A follow up on story 5: We really think adding of software executable under the software package is needed. Here's our use case.. Recent development of Grisu ( http://www.grid.apac.edu.au/repository/trac/grisu/) has shown that a *Generic Submission Tool* is possible and would be very useful. A software developer at a site could publish information for a new application that would then be instantly available for use. No generic tool can work without being able to query for the executable to use. Use case (for GLUE1.3 comment/issue #1): A user needs to know how many CPUs he can use for his MPI job. - FreeJobSlots which is the number of single processor jobs I can submit to a queue doesn't really answer this requirement. A queuing system reporting 32 free job slots wouldn't necessarily mean that the user can run a 32 CPU mpi job on that system. I've noticed the addition of the MaxSlotsPerJob attribute in the ComputingSharePolicy which you probably were thinking of using to address this use case. The use of this MaxSlotsPerJob term makes FreeJobSlots more confusing. If FreeJobSlots equates to the number of sequential jobs I can run on the system, I can't see any reason why a sequential job will have a limit to the number of processors it can use. As what we've talked about before in the telecon on 22 January, we need to have a new term to describe the free CPUs for the MPI job, FreeParallelCPUs perhaps?. And the term MaxSlotsPerJob can probably be renamed to MaxCPUsPerParallelJob. A comment on Stephen Burke's comment about GLUE 1.3 Issues/Comments #3: We don't think there's a way to calculate WorstResponseTime on a preemptive system. A worst response time for a job that user A submits now might not be valid anymore 5 minutes later because user B who has a higher priority submits a job and preempts the job submitted by user A. Use cases (for GLUE1.3 comment/issue #5 and #6): none - They were just comments and we don't really have use cases for them. Thanks, Gerson On Jan 23, 2008 3:50 AM, Sergio Andreozzi <sergio.andreozzi@cnaf.infn.it> wrote:
Dear all,
in the following page you can find links to the minutes of the two today telecons:
http://forge.ogf.org/sf/wiki/do/viewPage/projects.glue-wg/wiki/PhoneMeeting2...
the discussion on Storage entities will be continued Fri 25th 3 PM GMT.
The next telecon is for tomorrow and we'll resume discussion on policy and values when unknown.
http://forge.ogf.org/sf/wiki/do/viewPage/projects.glue-wg/wiki/PhoneMeeting2...
Cheers, Sergio
-- Sergio Andreozzi INFN-CNAF, Tel: +39 051 609 2860 Viale Berti Pichat, 6/2 Fax: +39 051 609 2746 40126 Bologna (Italy) Web: http://www.cnaf.infn.it/~andreozzi<http://www.cnaf.infn.it/%7Eandreozzi>
_______________________________________________ glue-wg mailing list glue-wg@ogf.org http://www.ogf.org/mailman/listinfo/glue-wg
glue-wg-bounces@ogf.org [mailto:glue-wg-bounces@ogf.org] On Behalf Of Gerson Galang said:
As what we've talked about before in the telecon on 22 January, we need to have a new term to describe the free CPUs for the MPI job, FreeParallelCPUs perhaps?.
A comment on Stephen Burke's comment about GLUE 1.3 Issues/Comments #3: We don't think there's a way to calculate WorstResponseTime on a
How does this work? I would have assumed that if a system allows MPI it would allow it on any CPUs. Are you saying that if a system has, say, 16 free CPUs and will allow up to 12-processor MPI jobs it may still be the case that a 12-processor job won't run because only 10 of the free CPUs are free for parallel jobs? preemptive system. A worst response time for a job that user A submits now might not be valid anymore 5 minutes later because user B who has a higher priority submits a job and preempts the job submitted by user A. I still think this is a misunderstanding. Even on non-pre-emptive systems the values may be wrong in 5 minutes time because more jobs may be queued in that time and they may go in the queue in front of yours, the published values are only ever a current snapshot. (Conversely, jobs in front may be cancelled or executing jobs may abort so the times may also suddenly get shorter.) Stephen
participants (3)
-
Burke, S (Stephen) -
Gerson Galang -
Sergio Andreozzi