Core Minutes 17 April 2007

Core Minutes 4/17/2007

Present: Joanne Bogart, Anders Borgland, Toby Burnett, David Chamont, Eric Charles, Jim Chiang, Richard Dubois, Warren Focke, Tom Glanzman, Berrie Giebels, Navid Golpayegani, Heather Kelly, Michael Kuss, Bryson Lee, Chuck Patterson, Tracy Usher

GR status, 55 day run: (Richard) We were generating events over the weekend until Bill called a halt. He had realized that the merit variables for ACD energy were based wholly on MC truth—not acceptable. Eric fixed it on Monday. Toby finished the all-gamma runs he was doing at UW (50 million events). We also have a cosmic muon run and a 1-day sample run. These are still useful for pass 4. We'll redo with GR v10r7 (containing Eric's fix). See latest information on the 55-day run checklist page.
We are about to see a request for a pointed run with the LAT pointed "backwards" (towards the earth). We'll do a 1-day sample run this way. (Toby) For this run spacecraft geometry is particularly important. (Richard) True. [Post-meeting Joanne suggested making another GR tag, differing from v10r7 only in the xmlGeoDbs tag, so we can make some direct comparisons.]

(Toby) Shall we go ahead with the new slimmed-down Gaudi? (Heather) Should be fine. No problems have been seen in LATEST.

(Jim) Seth reports that the latest version of Pulsar (v2r3p3) takes an inordinately long time to run using the Service Challenge pulsar model. Max Razzano has been notified and is looking in to the problem.
ROOT meeting: (David) See his report concerning subjects of interest to us. (Heather) recommends moving to 5.14E (we're at 5.14D now) to pick up the Hadd fix.
(David) would like to validate the Skimmer with ROOT 5.14. EM is still at ROOT 5.10. When it's upgraded, will need to handle merging of files from the two different releases.

(Warren) Currently it takes hours to merge data—usable but very slow. Ultimately will need something better. (Anders) We can't do timing tests for Pipeline 2 without faster merging.
CMT Action Committee report (Heather) CMT v1r18p20061003 solves many, but not quite all, of the problems people have experienced with (non-component) shareables on Windows. The one hold-out is the CalibData; We can continue using the old, mildly unpleasant work-around. The new version of CMT is now available from the Installer. Note: MRvcmt users should upgrade to at least v0r22p2; v0r24p2 is recommended. (Jim) would like to see RM builds with the new CMT to make sure test programs run. What happened to plan of using it for LATEST builds? (Navid) didn't realize this had been decided. He will enable it.
RHEL4 (Heather) It's coming along. There are just a handful of packages left that don't build.

minutes index

J. Bogart, Last Modified: 01-Jun-2010 15:47:07 -0700