[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] out-of-memory issues in parallel universe
- Date: Wed, 19 Mar 2008 12:57:05 -0400
- From: "Robert E. Parrott" <parrott@xxxxxxxxxxxxxxxx>
- Subject: Re: [Condor-users] out-of-memory issues in parallel universe
I understand this solution, but not all my users do :->
As I understand your response, these properties will be considered for
all nodes on which the job is run ... is that the case?
I'm also/instead looking for a solution to enforce memory limits at
It looks as if a USER_JOB_WRAPPER with a ulimit line is the solution
here. Does that jibe with what others have done?
On Mar 17, 2008, at 11:36 AM, Greg Thain wrote:
Is there some way of specifying the image size, and restricting jobs
to larger memory compute nodes, for MPI jobs submitted in the
By default, Condor tries to run jobs only on machines that have enough
memory. Condor_submit does this by sticking the clause:
((Memory * 1024) >= ImageSize)
into the job's requirements. The problem is that Condor doesn't
priori how much memory the job will need (the ImageSize). So, it
an initial guess based on the size of the executable. This guess is
almost always wrong, almost always too small. If you have a better
guess as to the image size, you can put it in the submit file:
image_size = some_value_in_kbytes
And Condor will only match the job to machines (or slots) with at
that amount of memory.
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx
You can also unsubscribe by visiting
The archives can be found at:
Robert E. Parrott, Ph.D. (Phys. '06)
Associate Director, Grid and
Project Manager, CrimsonGrid Initiative
Harvard University Sch. of Eng. and App. Sci.
33 Oxford St.
Cambridge, MA 02138