ScienceTools: Seth presented the report. Highlights include a long list of supported platforms from FSSC and Jim's Likelihood Usage Notes.
FSSC Report: (Eric) Quiet.
What is where: (Richard) Because of miscommunication the Aug 4 - Sep 2 data as reprocessed for new alignment was not available from the astro server when people were assuming it was there. We need a more formal mechanism to keep track. We're about to embark on reprocessing merit and FT1.
GlastRelease: (Richard) Overlay work is done. Johann is getting all-gamma datasets ready for IRF generation; background runs will be next. (Leon) But something terrible has happened in the last day or two: the all-gamma runs with GR v17r4 are taking about an hour longer than they used to with v17r3. The ACD overlay algorithm is the culprit, but the two changes made there would seem to be innocuous.
(Leon) Concerning background runs: it used to be that the orbit didn't matter, but for overlays it does. The background runs need the same 53-day orbit as the all-gammas.
More merit variables (Tracy) We need more information to keep track of interacting particles. What we have now (start and end) is not sufficient. No plan yet, but he hopes to discuss it next week in Hiroshima. (Philippe) We also need the total mc hit information for ALL 3 subsystems. Otherwise it seems that we can't correctly compute the effective area.
Disk space: (Tom) u35 [our "new" 4.6 Mbyte build disk] is nearly ready to go, but not yet cleaned out. He needs to confirm that the files can all be erased without complicating the cleanup of the pipeline and datacatalog databases, and xroot directories. He has been distracted by the I & T clean-up. The latter is motivated by the need for more disk space for xrootd. The I & T clean-up could yield as much as 30 Tbyte, but we have a contractual obligation to keep that data accessible. We're now in the process of archiving it to HPSS. (Richard) Anders is finding out who might conceivably need to look at this data; (Anders) believes no one has looked at it for quite a while.
Documentation: (Chuck) sends this report:
In the routine updates category, I've added a page on the "Interactive LAT Source Catalog" based on input from Elisabetta Cavazzuti. It's linked in from the top right corner of the LAT Science Analysis "splash" page.
In the 'new' category, I'm continuing to work on the ACD, TKR, and CAL documentation and am now doing more writing than researching. If all goes well, I hope to have a rough draft on the processing of real data ready for a preliminary review/edit by SMEs by the end of the collaboration meeting, or shortly thereafter.
The basic approach is to divide the GLEAM application into the processing of real data and simulated data. There's still quite a bit of work on this one, but I now have good traction, and the basic organization and structure is pretty much in place, though it will continue to evolve as we fill in the gaps.
HEAdas (Heather) will chat with Tom about using the Kipac Software collection. Specifically we need to determine how best to create any necessary set up scripts Tom will need for the Pipeline. Jim has been in touch with Stuart Marshall one of the maintainers of the software collection, and has reached an agreement where we can make use of the software where it is currently installed. If they find our Pipeline use of the software has a negative impact on their load, it will be addressed in coordination with SCCS at that time.
ROOT Compatibility (Heather) writes
Questions remain concerning skimming merit files written with ROOT 5.20. There are multiple reports of the skimmer built against 5.18e failing when used to skim ntuples created with 5.20. I did track down the issue to a problem in ROOT itself, and not the skimmer. A query has been sent to ROOT-central. This issue impacts the upcoming reprocessing, where only the merit files will be re-generated. The full ROOT files (digi, recon, relation) were written using ROOT 5.18c. This may complicate the use of the skimmer, since we would prefer to use the xxxRootData libraries built against ROOT 5.18c to handle the full ROOT files. More to come....
ACD updates (Heather) Eric Charles' updates are pending; nothing in CVS yet.
RM (Navid) Windows RM is now up and running. There is a problem with submitting to the RHEL5 machines. Mac builds via lsf are still down; he can make them manually via ssh.
SCons & Windows (Joanne) SCons (vs_revamp branch with additions from Tracy) is making project and solution files, but Express 2008 is not happy with them. I installed Professional 2008 overnight. It might not make the problem go away — most likely won't, in fact — but it's worth a shot, especially since that's the platform we really need to support, not Express. See also latest entries in the log.
ROOT Mac build (Emmanuel) reports:
After the formal build of ROOT 5.20.00 on the Mac, I then rebuilt pyROOT as a bundle instead of a shared library. This allowed me to overcome the python error, "Fatal Python error: Interpreter not initialized. (version mismatch)," when the pyroot demo.py attempted to import libPyROOT.so. I proceeded with testing ROOT via running, "demo.py," but I received several errors before receiving the prompt, "Enter q to quit" while running the demo. I am currently working on resolving those errors.
previous | minutes index | next |