[Yt-dev] RMS mass overdensity method
Stephen Skory
stephenskory at yahoo.com
Thu Mar 5 13:31:27 PST 2009
Matt,
> I'll be gone all day tomorrow until the evening driving to UCSB, but
> I'll be back late at night and maybe for a little while in the early
> or late afternoon. Let us know how Kraken goes, because this is
> confusing to me. Confusing and infuriating.
I hope your drive went well, as I assume you'll be done with it when you get this. I enjoy driving on US-101 between San Jose and Santa Barbara, at least compared to I-5.
I am having trouble running python on the compute notes on Kraken. I'm in a debug job: "qsub -I -V -q debug -l size=8;walltime=2:00:00." I can get the static version of python run, but not on the compute nodes. You can see that it's not related to the search path. See below:
http://paste.enzotools.org/show/76/
I compared 'printenv' and 'aprun printenv' and they are identical, so it's not an environemtal issue. Maybe I did something wrong with my build process? It follows the instructions we put together. And I'm pretty sure that printenv gets run on the compute node:
sskory at kraken-pwd3(XT5):~/l/voboz64/DD0130> hostname
kraken-pwd3
sskory at kraken-pwd3(XT5):~/l/voboz64/DD0130> aprun hostname
nid14592
Application 24719 resources: utime 0, stime 0
Am I trying to run this in a different way than you did, Matt? I should note that trying 'aprun mpi4py' gives nothing different.
Thanks!
_______________________________________________________
sskory at physics.ucsd.edu o__ Stephen Skory
http://physics.ucsd.edu/~sskory/ _.>/ _Graduate Student
________________________________(_)_\(_)_______________
More information about the yt-dev
mailing list