Difference between revisions of "Staff Priorities - Scott"
From SCECpedia
Jump to navigationJump to search| Line 1: | Line 1: | ||
== Yesterday == | == Yesterday == | ||
| − | * | + | *Tested buffering code, got better performance: |
| − | * | + | **read_sgt decreased from 21.1 avg to 15.8; max from 5322 to 1404 |
| − | * | + | **sgt_subset improved but could still be better: avg from 1327 to 664, max from 6773 to 5443 |
| + | **overall down to 5400 SUs/run | ||
| + | *Corresponded with Maren and David about generating seismogram data products for EEW | ||
== Today == | == Today == | ||
| − | * | + | *Testing 2 simultaneous workflows for SU burn |
| − | * | + | *Email Lonnie Crosby about Kraken I/O |
| + | *Investigate changing order of points in SGT | ||
*Work on new PGI seg fault | *Work on new PGI seg fault | ||
*ucvm2query | *ucvm2query | ||
Revision as of 18:08, 5 November 2012
Yesterday
- Tested buffering code, got better performance:
- read_sgt decreased from 21.1 avg to 15.8; max from 5322 to 1404
- sgt_subset improved but could still be better: avg from 1327 to 664, max from 6773 to 5443
- overall down to 5400 SUs/run
- Corresponded with Maren and David about generating seismogram data products for EEW
Today
- Testing 2 simultaneous workflows for SU burn
- Email Lonnie Crosby about Kraken I/O
- Investigate changing order of points in SGT
- Work on new PGI seg fault
- ucvm2query
Blocked
- No
Follow-ups
- No
- Run test of I/O forwarding on Kraken
- Continue testing compiler optimizations
- Run some I/O tests to get reference values
Areas of Responsibilities
- CyberShake 1.4
- Broadband CyberShake
- Code migration
