Difference between revisions of "Euler III Beta Testing"

From ScientificComputing
Jump to: navigation, search
(Clarified wording.)
(Parallel jobs)
Line 12: Line 12:
 
  bsub -R beta [other bsub options] ./my_command
 
  bsub -R beta [other bsub options] ./my_command
 
''Only'' jobs from approved beta testers will run as long as the system remains in this stage of the beta test.
 
''Only'' jobs from approved beta testers will run as long as the system remains in this stage of the beta test.
 +
 +
=== Submitting parallel jobs ==
 +
 +
While the Euler III nodes are targeted to serial and shared-memory parallel jobs, multi-node parallel jobs are still accepted. You need to tell the system that Infiniband is not available,
 +
module load interconnect/ethernet
 +
''before'' loading any MPI module. Also you need to request four cores per node:
 +
bsub -R beta -R "span[ptile=4]" [other bsub options] ./my_command
 +
Open MPI 1.6.5 has been tested to work. MVAPICH2 does not work.

Revision as of 14:22, 19 January 2017

The Euler III extension to the Euler cluster is available for beta testers.

If you are interested in running test jobs on the new nodes and are willing to work with us to diagnose and resolve problems, then contact us to be granted access as a beta tester for these nodes.

Serial jobs or single-node parallel jobs using from 1 to 4 cores, use up to 30 GB of total memory and request up to 24 hours are good candidates to run on these nodes.

The new nodes run CentOS 7 exclusively. All other production nodes in Euler currently run CentOS 6. Jobs that rely on Infiniband will not run on these nodes. Single-node Open MPI jobs will run but MVAPICH2 jobs will not run.

Submitting beta jobs

To submit a job to run on the beta Euler III nodes, you must request the beta resource, e.g.,

bsub -R beta [other bsub options] ./my_command

Only jobs from approved beta testers will run as long as the system remains in this stage of the beta test.

= Submitting parallel jobs

While the Euler III nodes are targeted to serial and shared-memory parallel jobs, multi-node parallel jobs are still accepted. You need to tell the system that Infiniband is not available,

module load interconnect/ethernet

before loading any MPI module. Also you need to request four cores per node:

bsub -R beta -R "span[ptile=4]" [other bsub options] ./my_command

Open MPI 1.6.5 has been tested to work. MVAPICH2 does not work.