Staff Priorities - Scott: Difference between revisions

From SCECpedia
Jump to navigationJump to search
Scottcal (talk | contribs)
No edit summary
Scottcal (talk | contribs)
No edit summary
Line 1: Line 1:
== Yesterday ==
== Yesterday ==
*Ran fwd test of AWP and RWG code
*Looked at Shiyu's volume results, 1D results - emailed out to group
*Tested a hunch about mesh inversion - success!
*Submitted TAU profiling run on Kraken
*Asked Shiyu to adjust AWP-ODC-SGT code and run again
*Identified 3.5 TB of SGTs to move to Ranch, started transfer
*Tested MPI I/O version of SGT Merge - matched with 1 core
*Continued work on RupGen-api memcached support
*Continuing production runs - 180k SUs on Ranger


== Today ==
== Today ==
*Check Shiyu's adjusted results
*Ask Shiyu to generate SGTs for volume
*Set up next SGT runs in 3D velocity volume (save volume points)
*Check on TAU profiling results
*Test SGT Merge with multiple cores on hpc
*Monitor Ranch transfer and md5sums (very slow!)
*Test SGT Merge on Kraken
*Test SGT Merge on Kraken
*Add memcached to in-memory RV generation
*Continue work on RupGen-api memcached support


== Blocked ==
== Blocked ==
*No, but wish USC webmail would work...
*No


== Follow-ups ==
== Follow-ups ==
*ORNL call (helpful)
*No
*CyberShake call


== Areas of Responsibilities ==
== Areas of Responsibilities ==

Revision as of 18:02, 2 March 2012

Yesterday

  • Looked at Shiyu's volume results, 1D results - emailed out to group
  • Submitted TAU profiling run on Kraken
  • Identified 3.5 TB of SGTs to move to Ranch, started transfer
  • Continued work on RupGen-api memcached support

Today

  • Ask Shiyu to generate SGTs for volume
  • Check on TAU profiling results
  • Monitor Ranch transfer and md5sums (very slow!)
  • Test SGT Merge on Kraken
  • Continue work on RupGen-api memcached support

Blocked

  • No

Follow-ups

  • No

Areas of Responsibilities

  • CyberShake 1.4
  • Broadband CyberShake
  • Code migration