This is a static archive of the previous Open Grid Forum GridForge content management system saved from host forge.ogf.org file /sf/wiki/do/viewPage/projects.glue-wg/wiki/PhoneMeeting20080122Part1Minutes at Sun, 06 Nov 2022 11:42:27 GMT SourceForge : View Wiki Page: PhoneMeeting20080122Part1Minutes

Project Home

Tracker

Documents

Tasks

Source Code

Discussions

File Releases

Wiki

Project Admin
Search Wiki Pages Project: GLUE     Wiki > PhoneMeeting20080122Part1Minutes > View Wiki Page
wiki1928: PhoneMeeting20080122Part1Minutes

Participants

  • Sergio Andreozzi (CNAF)
  • Laurence Field (CERN)
  • Felix Ehm (CERN)
  • Gerson Garlang (GridAustralia)
  • Stephen Burke (RAL)
  • Balazs Konya (Lund University)

Minutes (draft)

Action Revision

  • talk to JP and verify sc5 proposed solution and their use case;
  • check if WN environment work doc defined by GIN solves the problem (seems no because you need the info at scheduling time)
  • check if sc6 is fine with JP
  • check sc7/9/10/11 with JP

Gerson emailed to JP, expressed interested for ApplicationEnv entity

  • about executable, they use a different approach but this can be useful
  • they naming conventions (to ask for)
  • ACL not yet published by TeraGrid but can be useful

According to enumerations these are ok:

- Discussion on GLUE 1.3 issues/comments

1. CPU vs JobSlot: clarification why we moved away from FreeCPUs towards FreeJobSlots Action:

  • Gerson to send use cases for multi-processor jobs
  • Stephen to ask Cal Loomis for EGEE and people from GridIreland
  • Balazs: what can we expect that need to be changed to consider multi-processor jobs?
  • what about virtual machines and execution enviroment
Action: collect use cases
    • Balazs for NorduGrid
    • Sergio for INFN Tier 1
    • to dedicate a whole telecon on this topic
  • Laurence: some siteAdmin specify maxMemory usable by the job in the LRMS;

Action: collect use cases about different ways to configure per-job memory limits on batch systems

  • Balazs
  • Laurence to talk to Steve Traylen
  • Sergio to talk to INFN T1 people
  • Gerson to produce examples about physical/logical CPUs in real systems in AustralianGrid and how they are allocated to jobs

sc14: for parallel/sequential execution, probably we need two attributes:

Parallel: boolean (if false, only seq execution is supported) ParallelType: enum: mpi, opennmp, ...

accepted

after collection of use cases, make a final call to validate all the choices

 




The Open Grid Forum Contact Webmaster | Report a problem | GridForge Help
This is a static archive of the previous Open Grid Forum GridForge content management system saved from host forge.ogf.org file /sf/wiki/do/viewPage/projects.glue-wg/wiki/PhoneMeeting20080122Part1Minutes at Sun, 06 Nov 2022 11:42:42 GMT