Difference between revisions of "Staff Priorities - Scott"

Revision as of 17:17, 24 April 2012

Found more memory issues on Kraken. Explored and discovered some of the SRFs are corrupt (hence asking for weird quantities of memory)
Started Globus Online job to transfer corrupt files (based on md5)
Calculated hazard curve with AWP SGTs - lower (high-PSA variations lower by 4x). Running some seismograms by hand.

@@ Line 1: / Line 1: @@
 == Yesterday ==
-*Added CyberShake section to pegasus-mpi-cluster XSEDE paper
+*Found more memory issues on Kraken.  Explored and discovered some of the SRFs are corrupt (hence asking for weird quantities of memory)
-*Repeated RWG/AWP comparison with CVM-H, looks good
+*Started Globus Online job to transfer corrupt files (based on md5)
-*Reran AWP with CVM-S TEST site with modified Q - made mistake, running again
+*Calculated hazard curve with AWP SGTs - lower (high-PSA variations lower by 4x).  Running some seismograms by hand.
-*Checked on new TAU run on Kraken - need to rerun
-*Monitor workflow run
-*Talked to Gideon - killing workflow, need to change cores/node
 == Today ==
-*Working on correctly specifying cores/node
+*When SRF staging completes, rerun workflow on Kraken
-*Add references to CyberShake section in paper
+*Run seismograms by hand with AWP and RWG, investigate big differences
-*Look at plots from AWP forward run if new CVM-S run still has ringing
+*Continue development of JBSim3d + memcached + in-mem RV + PSA calculation
-*Add default striping change on Kraken dirs to workflow
+*Ponder way to reduce number of seismogram files
 == Blocked ==