Core Minutes 4 May 2010

Core Minutes 5/4/2010

Present: Joanne Bogart, Anders Borgland, Jim Chiang, Richard Dubois, Warren Focke, Tom Glanzman, Elizabeth Hays, Julie McEnery, Chuck Patterson, Leon Rochester, Tom Stephens, John Vernaleo, Eric Winter

ScienceTools: (Jim) See the Science Tools blog in Confluence for the latest news.
FSSC: (John) is ready for dual ingest. It's now up to SLAC to produce test files.
Documentation: (Chuck) sends the following:

Updates to gtgrb, the GRBanalysis, tutorial are pretty much complete; I'm making a final pass this afternoon and will then post it to the web. This version of the tutorial assumes that an SCons build of ScienceTools is used, and everything works with no problem except for XSPEC. XSPEC fails due to a glitch in gtgrb itself, which Nicola has already fixed in the HEAD build. This fix will be included in the next gtgrb release, and the tutorial documents that fact. I've removed the CMT version of the tutorial, and have not tested it with the ScienceTools-v9r16p01 build. I did see that in Nicola's gtgrb_slac script (in which the CMT section is commented out), there is a statement that it does not work. I'll revisit this when time permits to determine whether his note refers to the CMT build, or to the environment variables set up by the script.
GRB: (Richard) A new check-out package for GRB is in the works.
Pass 7 plans: (Anders) Making progress [but Anders himself, or at least his audio, faded away]

(Leon) The skim for overlays has severe memory problems. Even a simple job which does nothing but read digis has a big leak. This job has been examined with ValGrind, which shows that the leak is coming from TkrDigi (or something that TkrDigi uses). This is a problem we've seen before and fixed; the recurrence is perhaps due to the new ROOT. Once it's tracked down and fixed, we'll go back to the original job and see if the other leaks—we know there are others besides the one in TkrDigi—have the same cause.

[Wednesday update: Tracy has found and fixed the problem. The different leaks observed have the same cause. It's not in ROOT itself nor in TkrDigi, but between the two in RootIo.]
Pass 8: (Leon) Tracy reported results for global pattern recognition. There are still some loose ends but it's close to done and looks promising. It's somewhat faster than the production pattern recognition.

(Leon) Luca Baldini and Johan are working on error estimation for Cal clusters. They're looking into using methods developed originally for problems in geophysics.

minutes index

J. Bogart, Last Modified: 01-Jun-2010 15:47:14 -0700