Opened 11 years ago

Last modified 11 years ago

#816 new defect

IZ3279: Description of "job_is_first_task" in `man sge_pe` should be rephrased

Reported by: reuti Owned by:
Priority: normal Milestone:
Component: sge Version: 6.2u5
Severity: Keywords: man


[Imported from gridengine issuezilla]

        Issue #:      3279             Platform:     All      Reporter: reuti (reuti)
       Component:     gridengine          OS:        All
     Subcomponent:    man              Version:      6.2u5       CC:    None defined
        Status:       NEW              Priority:     P3
      Resolution:                     Issue type:    DEFECT
                                   Target milestone: ---
      Assigned to:    andreas (andreas)
      QA Contact:     andreas
       * Summary:     Description of "job_is_first_task" in `man sge_pe` should be rephrased
   Status whiteboard:

     Issue 3279 blocks:
   Votes for issue 3279:

   Opened: Tue Aug 10 13:13:00 -0700 2010 

The man page of sge_pe mentions for "job_is_first_task":

"..., while a value of FALSE indicates that the job script (and its child  processes)  is
not  part of the parallel program (the number of slots reserved for the job is the number of slots requested with the -pe
switch + 1)."

This is not what is happening. SGE will never lower the available slot count by 5 when only 4 slots are requested. Instead the parameter
controls, whether a local `qrsh` is allowed N or (N-1) times, when N is the number of granted slots on the master node of the parallel job.

(We don't make a sophisticated accounting. But  the following paragraph stating "...accounting information gets multiplied by number of
slots + 1" may also not be correct. When the job script is nothing doing at all besides having one `mpiexec` and only all spread slaves are
processing while the `mpiexec` is waiting, why should the reserved usage be multiplied by slots + 1? Is this really happening?)

Change History (0)

Note: See TracTickets for help on using tickets.