Difference between revisions of "Staff Priorities - Fabio"
From SCECpedia
Jump to navigationJump to search| Line 1: | Line 1: | ||
== Yesterday == | == Yesterday == | ||
| − | * | + | * Song RMG |
| − | * | + | ** Reviewed Part-B results on BBP server |
| − | * | + | ** Reviewed Part-B results from HPCC |
| − | * | + | ** Mw 5.5s ran fine, other crashed with memory errors |
| − | * | + | ** DLEN=DWID=100 meters causes memory usage to climb to 30G |
| − | * | + | ** Re-ran with 500m, then 400m, then 300m DLEN, DWID |
| + | ** 300m starts to fail --> try reducing the number of jobs per node | ||
| + | * Contacted KO, RT regarding BBToolbox version | ||
== Today == | == Today == | ||
| − | * | + | * Integrate latest BBToolbox version into BBP |
| − | * | + | * Unit test for RMG |
| − | * | + | * Check-in updated cluster scripts into svn |
| − | * | + | * Update trac tickets |
| + | ** Add -ppn option to cluster scripts | ||
== Blocked: == | == Blocked: == | ||
| − | * | + | * HPCC Down: Cannot continue validation of Song RMG |
| + | * UCVM etree: Need David to help debug sort/merge issue on BW | ||
== Followups: == | == Followups: == | ||
Revision as of 17:34, 12 October 2015
Contents
Yesterday
- Song RMG
- Reviewed Part-B results on BBP server
- Reviewed Part-B results from HPCC
- Mw 5.5s ran fine, other crashed with memory errors
- DLEN=DWID=100 meters causes memory usage to climb to 30G
- Re-ran with 500m, then 400m, then 300m DLEN, DWID
- 300m starts to fail --> try reducing the number of jobs per node
- Contacted KO, RT regarding BBToolbox version
Today
- Integrate latest BBToolbox version into BBP
- Unit test for RMG
- Check-in updated cluster scripts into svn
- Update trac tickets
- Add -ppn option to cluster scripts
Blocked:
- HPCC Down: Cannot continue validation of Song RMG
- UCVM etree: Need David to help debug sort/merge issue on BW
Followups:
- No
To Do:
Short Term
Broadband Platform
- Update GP rupture generator - Total: 3 days
- Sent Rob parameters - done
- Integrate code - 1 day
- Run tests for Part-A / Part-B - 2 days
- Anderson GoF code from AS - Total: 3 days
- Analyze code, dependencies, provide AS feedback: 1 day
- Convert code to BBP module, add module to BBP workflow: 1 day
- Test BBP platform, produce new metrics: 1 day
- New BBToolbox version available from SDSU - Total: 3 days
- Incorporate SDSU changes into BBToolbox: 1 day
- Run Part-A/Part-B validations: 2 days
- NGA-West 2 validation on BBP - Total: 6 days
- Contact CG, collect input scenarios: 1 day
- Write test code: 1 day
- Run tests: 2 days
- Review results: 2 days
- Setup laptop as EEW machine in software lab - Total: 1 day
- Erase laptop, install clean Mac OS X, install EEW software: 1 day
- Integrate Rupture Generator code from SG - Total: 28 days
- Review code, understand inputs/outputs, run outside platform: 1 day
- Run Part-A problem using Matlab code, review results: 1 day
- Migrate code into BBP Platform, convert Matlab code into Python/C/Fortran, test: 20 days
- Run Part-A/Part-B validation exercise on HPCC: 5 days (contingent on code performance)
High-F
- Set up High-F code repository - Total: 3 days
- Collect codes from groups, scripts - 1 day
- Select SVN/Github repository - 1 day
- Create repository - 1 day
- Understand how to map station location to grid point - Total: 2 days
- Understand approaches used by AWP and Hercules to map station location to grid point - 1 day
- Discuss with PM, DG, how to use UCVM to do this mapping - 1 day
Medium Term
- Package GF generation code from RG into BBP Platform
- New raytracing code for SDSU method
- Update Anderson method
- Try logging into Gordon
