Present: Toby Burnett, Xin Chen, Richard Dubois, Marco Frailis, Riccardo Giannitrapani, Berrie Giebels, Navid Golpayegani, Traudl Hansl-Kozanecka, Julie McEnery, Monica Pepe, Tracy Usher
ntupleWriterSvc: (Toby) Toby has updated this service to add a new named service which will take pointers to a list of doubles. He has also added the feature of directly writing Root ntuples (so they are not entirely lost if the job crashes). He has been awaiting Heather's approval.
GlastPolicy, GaudiPolicy: (Toby, Traudl) The cleanup continues, with much more documentation. The plan is to supply some overriding patterns so that new packages can apply a single pattern, depending on what package type they are.
Cygwin: (Toby) Toby is banking on the claim that the current cygwin is now a variant of linux, and could provide a means for Windows developers to have an embedded linux system on their machines without a dual boot. [Ed. might be handy for me since we were unable to dual boot my new Dell D600] But we'll have to see what the price tag of support is.
Sources: (Toby) flux is now back in its own package. Theodore is checking out Mizuno-san's CR flux source. They still have to iterate a bit on how the magnetic latitude is handled. For batch job management, FluxSvc allows specifying start and end times. Toby would like to bulk up GlastRandomSvc to allow seeding events by timestamp, and will work with Xin on it.
S-Plus: (Toby) A heads-up from Toby. He has been working with a summer student to implement what Bill is doing in Insightful Miner in S-Plus. This may also be reproducible in (free) R.
MC Pruning: (Tracy, Riccardo) Tracy and Riccardo have implemented a fix to the PruneCal mode wherein it uses volume id's rather than z=0 to eliminate shower particles originating and terminating in the CAL. Riccardo will also look at the default pruning to also handle EM processes other than pair conversion (if the process is not pair conversion, no daughters are kept at present).
FRED status: (Riccardo) Riccardo has FRED up and running on the UW terminal server. He does not have the corba part working on linux yet. And has not looked at running FRED via Gaudi from Root (a Gaudi-light version, with just RootIo).
Onboard Filter embedding: (Navid) David Wren has run off 100k events with no crashes. Navid is awaiting specs for the filter-generated tracks to put on the TDS.
Release Manager problems: (Richard, Navid) The RM is in a bit of a tizzy trying to run the unit tests for the Latest builds. Many of them fail saying they can't find the exe. We don't know why, and Alex is still away. If Navid runs the script by hand, it works fine; but apparently not in the cron job. Alex copied all the code from his personal space to the group account. Also, we had an incident of renaming a package that the Latest builds are using. We think Alex will have to clean up when he gets back. For information, the cron job to look at tags, HEAD and LATEST builds runs hourly now.
ROOT memory leak: (Ursula, Richard, post meeting) It took a fix from the Root team to the cvs version of Root, but the TProcessIds were in fact the 'leak'. Root was keeping a list of all pointers across branches and events. There is now a TChain member function to turn this off.
50M background run complete: (Richard) Done. Now to clean up, prune and point Bill at the added 16M events.
Action Items
previous
|
minutes index
|
next
|
R.Dubois Last Modified: 01-Jun-2010 15:46:26 -0700