[torqueusers] pbs_nodefile parallel job mother superior node listed position reliable?

Garrick Staples garrick at clusterresources.com
Wed Sep 27 11:49:46 MDT 2006


On Wed, Sep 27, 2006 at 06:31:11PM +0100, David Golden alleged:
> Somewhat silly question, but it suddenly began to trouble me:
> Is the first node listed in $PBS_NODEFILE in a parallel job *defined* to be 
> the "mother superior" for the job, or  does it just happen to work out that 
> way right now?
> 
> i.e. can I reasonably safely assume the first node will be the m.s. when 
> processing $PBS_NODEFILE within a job, or should I go do something more 
> involved to cross-check? 

Yup, that is safe.



More information about the torqueusers mailing list