Core Minutes 9/20/2005
Present: Joanne Bogart, Toby Burnett, David Chamont, Jim Chiang,
Richard Dubois, Dan Flath, Warren Focke, Berrie Giebels,
Tom Glanzman, Navid Golpayegani, Heather Kelly, Michael Kuss, Matt Langston,
Julie McEnery, Chuck Patterson, Igor Pavlin,
Leon Rochester, Tom Stephens, Tracy Usher
- Report from Udiné: According to Riccardo,
his actual condition.
- DC2 news:
- ACD geometry
(Richard) A test data set for a build with the new ACD geometry (shrunken
tiles and embedded screws) is now available for examination. No problems
have been seen so far, but it needs more scrutiny.
- Next up are higher Tracker cluster cut-off
(changed from 500 to 2000) and improved Onboard Filter
from OSU. Consensus at the meeting was to put both in together in the next
HEAD build; Richard is feeling lucky. As soon as it's built, Matt is poised
to make a test run with it.
- RootIO problem (Leon) Track ids are not getting written for
hits which are not in a cluster. This causes FRED some grief. The fix
(to fill in the id with something sensible) is straightforward, but might
lead to backwards-compatibility problems. (Heather) believes we could
most likely avoid such problems; she'll look into it.
- Pointing history (Richard, Toby) ..has not been getting put into
the merit tuple, but Toby has just put a fix into FluxSvc.
Some problems Leon has seen while engaged in a project to see if he can
generate the complete merit ntuple from the ROOT trees
may be related to this.
- Reproducibility (Michael) is checking this out for DC2. He had
some start-up problems, but is past most of them. He hopes to have some
results in the next day or two. The goal is to be able to select events
of interest, get sources for them and rerun reproducibly.
- Viewing events (Richard) The MC tree isn't (yet) terribly
photogenic; trajectory information is not normally included and by default
only the minimal particle tree is written. This is especially unsuitable
for heavy ion events. (Tracy) is looking into writing out more information,
at least for certain selected events. But, if we can reproduce the event
from the source, he wonders if this is really necessary. (Richard) We
can only hope to reproduce on the same OS.
- Billion schedule Bill suggests October 1 or thereabouts.
- Down on the (Lyon) farm
Berrie plans to make pretty much the same kinds of runs as will be done
at SLAC, other than scaling down the billion event run somewhat. Tom and
Matt will document procedures at SLAC so he can figure out how to adapt
to his environment. He will FTP his output to SLAC and Tony will catalog
it (Tony expects to have this ability later this week).
- Release Manager/GlastPolicy update: (Navid)
He's been fighting other fires recently, so status hasn't changed from
last week. The new GlastPolicy hasn't yet gone into a HEAD build, but is
ready to do so.
- Pipeline: (Richard) Tom has several
suggestions for improving the Pipeline. He, Dan, Tony and Matt will be
meeting later today to discuss them.
- Coming attractions: Next week Tom will talk
about his work with Andrey on using Python with Gaudi and ROOT.
J. Bogart, Last Modified:
01-Jun-2010 15:46:39 -0700