Core Minutes 4/18/2006
Present: Joanne Bogart,
Toby Burnett, Eric Charles,
Jim Chiang, Richard Dubois, Warren Focke,
Berrie Giebels, Tom Glanzman, Navid Golpayegani,
Heather Kelly, Michael Kuss, Bryson Lee, Francesco Longo,
Julie McEnery,
Chuck Patterson, Igor Pavlin, Leon Rochester, Tom Stephens,
Tracy Usher
- Processing Agenda(Richard) Cosmic ray
running is expected to start tonight and go for two weeks. This will
take up most of our processing power during the night. The rest of
the time can be used for finishing up DC2 and for Beamtest simulation.
- DC2: (Richard) Tom has successfully
rerun all the pulsar jobs since the problem was tracked down last week.
Seth has looked them over and is satisfied they're ok. Still need to
do post-processing, including
- Computing livetime (cumulatively over DC2, not just job by job)
- Add columns to ntuple for ACD variables Bill wants
- Add rescaled CTB-gamma. Julie clarified the fact that this is
not a replacement or redefinition of the existing number. It will be
an addition with a new name. It will have the definition Bill has been
using, which he
proposed before DC2 Kick-off; there just wasn't time then to
get it in before.
- Riccardo Rando's classification tree.
and then calculate diffuse response.
If all goes well this will be done by tomorrow and the new DC2 sky will
be available by the end of the week.
- Beamtest: (Francesco, Michael)
Some pipeline runs went well for the most part; there are some small
issues to be dealt with. Johann is working on
an ntuple, a modified version of the SVAC ntuple. The attempt to
build Beamtest06 against GlastRelease v9 went smoothly. [Nonetheless,
Richard says first runs will be made with current Beamtest (built
against GR v8 except for a geometry update). Tom is running
100k 2 GeV electrons as a test.]
(Richard)
What about SPS simulation. (Francesco) There is still a small amount
of work there. (Richard) It sounds like a couple more days before
you'll be ready for bulk simulations.
- Status of GlastRelease: (Toby)
Would like to get v9r1 out the door and look at background simulations,
but the process has been rocky. Developers should monitor RM
builds to confirm that their code works on both OSes, but this has
not been the case: there were several Windows-only problems.
In general we continue to suffer from lack
of Windows systems expertise.
(Heather) ..was remiss this time around because she thought it was being
dealt with, but in general is happy to keep an eye on Acd packages on
Windows. Things were complicated somewhat by the need for an I&T release,
so that much of the code in question was first implemented on a branch.
- Onboard Filter Toby with difficulty tracked down the
patch needed to avoid a divide-by-zero error. Somehow, although
the fix was in CVS, it was not visible in the CVS Changelog.
(Julie) There is another problem having to do with an array being
read in incorrectly which probably also needs a patch. She will
track it down.
- ACD Calibration code
There is one remaining problem having to do with parsing a calibration
file. (Eric) Actually the problem is that the file is not being found.
It needs to be copied over to the directory used by SLAC Windows
for calibrations. [Update: has been.]
(Leon) The unoffical convention has been that
subsystems take special action for flavor "ideal", avoiding the
need to access the MySQL
calibration metadata database and the normal pool of calibration files.
[Update: Eric took this on after the meeting;
it's about ready to go.]
[Update: The bent-tile geometry was inadvertantly
introduced into GR HEAD, but Acd code is not yet ready for it, so it has
been backed out. The geometry in GR v9r0 is fine for now.]
There was some question about System Tests for v9r0 waiting on an
optimized build, which in turn was waiting on a proper set of external
libraries. (Navid) The external libraries and a build using them are there.
Any remaining problems with the build have to do with the code, not
external libraries.
- Data handling: (Bryson) We're continuing with
database meetings, discussing trending databases and pointing history
among other things. There was some brainstorming last week on logging.
We plan to look into implementations used by other experiments, such as
Babar.
J. Bogart, Last Modified:
25-Jul-2007 11:25:29 -0700
previous
