Staff Priorities - Scott: Difference between revisions
From SCECpedia
Jump to navigationJump to search
No edit summary |
No edit summary |
||
| Line 14: | Line 14: | ||
*Finish interrupted PP runs on Ranger | *Finish interrupted PP runs on Ranger | ||
*Add memcached to in-memory RV generation | *Add memcached to in-memory RV generation | ||
*Try 27x weak scaling test on Kraken | |||
== Blocked == | == Blocked == | ||
Revision as of 18:23, 22 February 2012
Yesterday
- Compared SGTs from workflow SMALL site to by-hand SMALL site - looks fine
- Submitted run testing theory about rotation
- Tried Pegasus 4.0.0 with TEST curve, ran into issue, emailed Karan
- Ran TEST curve with in-mem RV generation on Ranger, passed
- Discovered slow performance of SGTMerge step with full-scale SGTs on Kraken
- Checked results from profiling test on Kraken - some kind of corruption in TAU files, emailed Matt
Today
- Adjusted SGT Merge parameters on Kraken, trying again - may need MPI I/O
- Email Rob F2F ruptures, ask for thoughts
- CyberShake call
- Check newest run results with 3D issue
- Finish interrupted PP runs on Ranger
- Add memcached to in-memory RV generation
- Try 27x weak scaling test on Kraken
Blocked
- No
Follow-ups
- CyberShake call
Areas of Responsibilities
- CyberShake 1.4
- Broadband CyberShake
- Code migration
