[torqueusers] why my job's resources_used.cput always 00:00:00

Tim Miller btmiller at helix.nih.gov
Mon Dec 18 09:18:35 MST 2006


luo yi wrote:
> Is there a bug in accounting  the resources of a job?  no matter how big 
> the job is ,the  resources_used.cput of the job always 00:00:00,and at 
> the same time the accounting of resources_used.walltime is pretty good. 
> this problem appears when i run Torque at a cluster which platform is 
> Redhat AS 4.0 and  nodes is 2-way  dual core xeon machine. but when i 
> run Torque at a cluster with Redhat 9.0 and 2-way sigal core xeon 
> machine,it can accounting the resources_used.cput correctly. why this 
> problem appears, how can i resolve it ????

Are you starting the jobs with mpirun? Torque can't do CPU time
accounting for mpirun started jobs. The solution is to use mpiexec or
just use the walltime (which is what I do).

Cheers,
Tim

-- 
Tim Miller
Contractor / System Administrator -- Laboratory of Computational Biology
National Institutes of Health   --   Bldg. 50 Rm. 3310  --  301-402-0618


More information about the torqueusers mailing list