March 8, 2012 Scalability Meeting
Date: March 8, 2012 Time: 10:00 AM EST
Call in information:
Toll Free (US): 888-426-6840
Toll Free (Germany): 0800-000-1018
Toll Free (UK): 0808-234-5071
- Greg Watson, IBM
- Dave Wootton, IBM
- Wolfgang Frings, Juelich
- Chris Navarro, NCSA
- Jeff Overbey, NCSA
- Jay Alameda, NCSA
- Claudia Knobloch, Juelich
- Actions from last meeting
- Greg and Chris resolved the SWT issue and closed the associated bug
- Greg still needs to talk to Jay
- Status Reports
- After examining the issues, it was decided to drop the NatTable approach in favor of continuing with the current custom table views. There are two bugs outstanding that need to be fixed.
- Now able to get usage data from GPUs using Torque. This will go into the next DA update, along with the node ordering bug fix.
- PBS BG/P support has been contributed by Paratools and is waiting for Eclipse Legal approval. Wolfgang to test out on his system.
- Contribution of generic SLURM and BQ/P and BG/Q support
- NCAR have LSF adapter for LLview. Need to obtain this to see if it can be adapted for LML. Dresden also have an LSF adapter we could look at.
- Updated PE job submission script
- Ran PE job on 3500 tasks over 29 nodes with no scaling issues.
- Continuing to work on bug that prevents output from being displayed when debugging a job.
- Actions Arising
- Greg to talk to Jay re funding
- Greg to enlist help on testing SLURM support
- Meeting adjourned; Next meeting March 15, 2012 @ 10am ET to be announced on ptp-scaling list. *Note DST change for USA*