[GE users] cputime for parallel jobs does not reflect the number of cpus used

Ron Chen ron_chen_123 at yahoo.com
Wed Mar 1 23:45:50 GMT 2006


    [ The following text is in the "iso-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

The parallel PE needs to be "tight" in order for job accounting
to work for slave tasks.

See "control_slaves" in the sge_pe(5) manpage:
http://gridengine.sunsource.net/nonav/source/browse/~checkout~/gridengine/doc/htmlman/htmlman5/sge_pe.html

Also take a look at the HOWTOs on the steps needed to get a
tight integration for your MPI library:
http://gridengine.sunsource.net/howto/howto.html#Tight%20Integration%20of%20Parallel%20Libraries

 -Ron


--- Nicolas Bock <nbock at lanl.gov> wrote:
> Hello list,
> 
> I was looking into how my resource usage changes after a job
> finishes  
> and noticed the following: When I submit a parallel job to the
> queue,  
> the cputime seems to only increment by the cputime used on the
> head  
> node of that job. It does not seem to increment by the sum of
> the  
> cputimes across all nodes. Is there a way to change that?
> 
> nick
> 
>
---------------------------------------------------------------------
> To unsubscribe, e-mail:
> users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail:
> users-help at gridengine.sunsource.net
> 
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list