Present: Joanne Bogart, Xin Chen, Richard Dubois, Heather Kelly, Michael Kuss, Sean Robinson, Leon Rochester, Alex Schlessinger, Karl Young
VRVS problems: Leon had trouble joining the meeting, but ultimately made it. Toby never did, so the meeting was short and sketchy.
Release Manager: (Alex) The summary of Release Manager status hasn't changed too much since last week.
Energy units: (Sean) Changes are in the works to require explicit energy units for sources. He is changing all the flux sources he is aware of, but people with private files will also have to make adjustments. To allow for this, there will be an announcement before the new code is incorporated.
(Joanne) Code changes to xmlUtil and detModel have been made to support a modified format for xxxResponse.xml files (part of detector description) containing constants in units of energy in which energy units (MeV or GeV) must be supplied. Output energies are always MeV. The changeover to XML files using the new format must be coordinated with ACD and CAL, preferably some time this week.
Gleam User's Guide: (Heather) In order to get it into shape for the Release, there are several items that ought to be addressed, including
provide an easy way to see all job options
document how to access TDS when using userAlg
ROOT classes: what's in them and how to access them
how to find out what sensible quantities to histogram might be. The cognoscenti browse in CVSweb. Would be nice to have a less intimidating method.
Thanks, Karl, for spending some time reviewing and volunteering. We need more volunteers to get this done. Start by looking at the current version of the Guide.
Test Releases: (Richard) Can't have much discussion of this issue without Toby, but the idea is to keep track of all packages of interest, whether or not they're currently used by Gleam. For each such package, keep track of most recent tagged version which is thought to be in good enough shape to be integrated with other such packages when building Gleam or other executables. These versions would then be the default to be used by nightly builds.
EXTLIB and access to external libraries: (Joanne) Right now we have no uniformity in the way interface packages are written. Also EXTLIB defines obscure macros like LHCBDBASE which it would be nice to get rid of. This should be cleaned up, though there is no particular need to do it in time for the Release. See this summary of the current situation. Further discussion put on hold until Toby can participate.
CAL leakage correction, memory: (Richard) In the course of looking into the CAL leakage correction discovered a significant memory leak, probably in showering. Jobs were run with 5, 10 and 50 GeV particles at several angles. All of the 50 GeV jobs crashed after working themselves up to 700 Mbytes. Even the 5 GeV jobs got up to 500 Mbytes. Also discovered that the leakage correction does need some work.
Core Week reviews aftermath: (Heather) We need minutes, reports on the two reviews which took place during Core Week.
previous | minutes index | next |
J. Bogart Last Modified: 01-Jun-2010 15:46:12 -0700