[Swift-user] control number of condor nodes?

[Allan Espinosa]

under the globus namespace:

<profile namespace="globus" key="spread">0.1</profile>

http://www.ci.uchicago.edu/swift/guides/userguide.php#profile.globus

2010/6/24 Glen Hocky <hockyg at uchicago.edu>:
> Which parameter is that?
>
> On Thu, Jun 24, 2010 at 11:40 AM, Allan Espinosa <aespinosa at cs.uchicago.edu>
> wrote:
>>
>> hi Glen,
>>
>> How about setting the spread to be more uniform (like 0.1)?
>>
>> -Allan
>> 2010/6/24 Glen Hocky <hockyg at uchicago.edu>:
>> > Hey Everyone,
>> > Quick question. I'm testing on bg/p and things are working fine, but I
>> > can't
>> > seem to increase the number of nodes I'm using in the way I expect. I
>> > can
>> > run fine on 64 nodes tih
>> >>
>> >>     <filesystem provider="local"/>
>> >>
>> >>     <execution provider="coaster" jobmanager="local:cobalt"/>
>> >>
>> >>     <profile namespace="globus" key="queue">prod-devel</profile>
>> >>
>> >>     <profile namespace="globus" key="slots">1</profile>
>> >>
>> >>     <profile namespace="globus" key="nodeGranularity">64</profile>
>> >>
>> >>     <profile namespace="globus" key="workersPerNode">4</profile>
>> >>
>> >>     <profile namespace="globus" key="maxNodes">64</profile>
>> >>
>> >>     <profile namespace="globus" key="project">OOPS</profile>
>> >>
>> >>     <profile namespace="globus" key="kernelprofile">zeptoos</profile>
>> >>
>> >>     <profile namespace="globus" key="maxtime">3559</profile>
>> >>
>> >>     <profile namespace="globus" key="alcfbgpnat">true</profile>
>> >>
>> >>     <profile namespace="karajan" key="jobThrottle">2.55</profile>
>> >>
>> >>     <profile namespace="karajan" key="initialScore">100000</profile>
>> >>
>> >>     <scratch>/tmp</scratch>
>> >>
>> >>
>> >>
>> >>  <workdirectory>/home/hockyg/reichman/glassy_dynamics/code/swift/work</workdirectory>
>> >
>> > I successfully went up to 256 nodes by increasing slots to 4 but I was
>> > unable to get a block of size 256.
>> > When I set maxnodes to 256, slots=1 , i get a block of size 128
>> > regardless
>> > of several changes (and removal) of jobThrottle
>> > Any ideas on how to fine tune this, because I will certainly need to
>> > choose
>> > my number of jobs based on an expected ( max wall time * workers per
>> > node *
>> > number of nodes )
>> > Glen
>
>



-- 
Allan M. Espinosa <http://amespinosa.wordpress.com>
PhD student, Computer Science
University of Chicago <http://people.cs.uchicago.edu/~aespinosa>