[torqueusers] Requesting memory as a resource with torque scheduler

Atwood, Robert C r.atwood at imperial.ac.uk
Thu Jun 14 09:20:19 MDT 2007


I have the same query but I do not see the answer to this on the list
(where the default scheduler is used). The documentation seems to
indicate that this should work, but none of my attempts to use it have
any effect. 

http://www.clusterresources.com/wiki/doku.php?id=torque:2.1_job_submissi
on

So can someone explain, preferably on-list? Should this be edited out of
the wiki, or do I not use it correctly?

Thanks
Robert


Here's an example, I also tried pmem, vmem or pvmem directives with no
observed effect. 


****** submitting a job requesting 26 giga when only 8 are installed and
4 swap
atwood:~/q> qsub qburn -lmem=26gb
11856.mt-xxxx.mt.ic.ac.uk 


****** the job is running on node01
qstat -an
...
11856.ic atwood short    qburn       17142     1  --   26gb 04:00 R   --
   node01/1


***** pbsnodes indicates total memory of 12 Gb as expected  
atwood:~/q> pbsnodes node01
node01
     state = free
     np = 2
     ntype = cluster
     jobs = 0/11855.mt-xxxx.mt.ic.ac.uk
     status = opsys=linux,uname=Linux node01
2.6.15.1-clustervision-81_cvos #1 SMP Thu May 18 10:47:02 CEST 2006
x86_64,sessions=17076,nsessions=1,nusers=1,idletime=2510121,totmem=12275
344kb,availmem=12066556kb,physmem=8178812kb,ncpus=2,loadave=1.30,netload
=49968056554,state=free,jobs=11855.mt-xxxx.mt.ic.ac.uk,rectime=118183405
5




> -----Original Message-----
> From: torqueusers-bounces at supercluster.org 
> [mailto:torqueusers-bounces at supercluster.org] On Behalf Of 
> Prakash Velayutham
> Sent: 30 May 2007 21:27
> To: torqueusers at supercluster.org
> Subject: [torqueusers] Requesting memory as a resource with 
> torque scheduler
> 
> Hello,
> 
> Is the following request supposed to work fine with Torque scheduler?
> 
> qsub -I -lnodes=1:ppn=4 -lmem=16000mb
> 
> I have several nodes with a total memory of 6 GB, some with 
> 12 GB, some 
> with 16 GB and some with 40 GB. But the above request always 
> puts me in 
> one of the nodes with 6 GB total memory (and available is less than 
> total, so no way it has 16 GB free). Can someone tell me what 
> is going 
> on here?
> 
> Thanks,
> Prakash
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
> 


More information about the torqueusers mailing list