[intrepid-notify] ALCF Weekly Updates - July 19, 2013
ALCF User Services
support at alcf.anl.gov
Fri Jul 19 12:57:41 CDT 2013
*** User Tip: Big Run Monday ***
If you are running capability jobs on Intrepid and want to reduce the
turnaround time of your simulations, remember to submit by Sunday since
every Monday, any jobs in the 'prod-capability' queue in the 'queued'
state will be promoted to highest priority. Visit the Job Scheduling
Policy on BG/P Systems webpage for additional information:
http://www.alcf.anl.gov/user-guides/job-scheduling-policy-bgp-systems.
*** Multiple Simultaneous Runs Within a Single Cobalt Job ***
It is possible to do simultaneous runs within a single Cobalt script job
with each run using a subset of the Cobalt job's nodes. For example, to
achieve capability levels you can bundle four 2K nodes job into a single
8K script job. This is accomplished with the command "cobalt-subrun",
which takes arguments in the same style as cobalt-mpirun, with the only
difference being the first two arguments have special meaning. See
examples of Cobalt script at
http://www.alcf.anl.gov/user-guides/bgp-running-jobs#advanced-job-patterns-using-scripts.
If you have any additional questions, contact support at alcf.anl.gov.
*** HPSS Data Archive and Retrieval System Available ***
The data file system on ALCF Resources is not backed up (e.g.,
/projects). It is your responsibility to ensure that copies of any
critical data on the data file system have either been archived to tape
or stored elsewhere. The HPSS data archive and retrieval system is
available on Mira and Intrepid to assist you in data management. For
more information on HPSS, please visit:
www.alcf.anl.gov/user-guides/using-hpss. To request access to HPSS,
please email support at alcf.anl.gov.
More information about the intrepid-notify
mailing list