In preparation of Richard's absence for hip surgery.
- SCons
Navid is chugging away and there do not seem to be any roadblocks at the moment to getting SCons and ST going. - Core Meeting early next year
Richard wondered if it would be advantageous to have another core meeting in the near future - when he is able to attend, of course. - EBF Blob for OktoberTest reprocessing
No fixed deadline for getting this done, but ASAP would be nice
Heather is peering into the evt and ebf blobs in HEX to see what may be different. - Interleave and CELs
Tracy sent mail to Richard stating that he has the CELs running. Tracy has sent mail to David for further clarification concerning how to best use the CELs with Interleave.
Heather will pay attention to see if any assistance is needed. - HADD Merge problems
Rene states that the problem merging the CAL and SVAC ntuples in this case is due to their sparseness. He suggests a different basket size or use the new THnSparse arrays. Richard wondered about the possibility of just using variable length arrays in the ntuples, despite the apparent desire on the part of users for fixed length arrays. - RHEL4
Zach is actively looking into the problem with rhel4_gcc34opt and the Gleam jobs, which seem to point to CalCalibSvc. The failures seem to be specific to rhel4-64.
Meanwhile, Tom and Richard have reported to SLAC Computing that the runtimes on the rhel4-64 machines are many times slower than would be anticipated. Babar has not yet migrated to rhel4 either, perhaps the will see similar processing speed issues. - GCC4 and 64 bit architectures and other compilers
How soon do we expect to move to gcc 4.0? Richard states that the Mac comes with gcc 4.0 by default. Also given our rhel4 adventures, and Zach's suggestion that the crashes only occur on rhel4-64 - do we need to consider porting our externals to rhel4-64? Should discuss with Navid about this possibility. Navid has stated in the past that such a move would require moving to current versions of our various externals. - JobOptions
The ongoing battle concerning the setting of TriggerAlg.applyPrescales is but one of the settings that can be different for MC, testing, or data processing. Eric Charles spoke at length during the last core meeting concerning a need for intelligent JO processing that could determine whether a run was MC or real data, and load the correct calibrations on the fly. Heather wants to look at the new Gaudi to see what if anything new is in the JobOptionsSvc. - OBF
Richard will send a note to Tracy, and then Julie and Heather should proceed to pester. Currently Tracy has OBF built. Anders and company are interested in the diagnostic filter, it is unclear what work is involved to get that going to their satisfaction. - Moot and Fun
Eric Grove had originally stated Moot and configuration handling needed to be in place by December. During the collaboration meeting, they backed off and said March would be ok. - MRvcmt & FRED support
Richard should speak to Francesco again, but at the moment the assumption is that someone will have to take over support for MRvcmt and Fred. Richard wondered aloud if NRL could be approached about allowing Zach to put some time into such infrastructure issues.