Core Minutes 11/2/2010
Core Minutes 11/2/2010Pass8: (Tracy) Runs take 10-20% longer than with our standard pattern recognition, but it's all due to 2 or 3 events per run.
Reprocessing: (Tom G.) Many thanks to Heather for finding and fixing the memory leak which has been plaguing us. (Heather) The leak occurred when reading recon ROOT data back into Gleam and was due to a bug in RootConvert which may have been there forever.
(Richard) When will L1 pick up the new Pass6 stuff? (Anders) I suppose when we are told to. C & A will decided.
[post-meeting note from Tom G.] I received an email from Sasha during our meeting that said he was happy with the 100 test runs and that we should go ahead with this reprocessing, the purpose of which is to create the missing GCR files between October 2009 and July 2010. This GCR reprocessing has now started.
Core talk:
Upgrade cfitsio? (Tom G.) The newer version does not fix the memory problem we were hoping it would address, but meanwhile gtsrcmaps magically fixed itself. The pressure to upgrade cfitsio is therefore off, but there are other reasons to prefer the new version. The version we're using is very old and there have been a number of improvements which look interesting.
Other externals (Tracy)
New Windows box (Richard) Wayne (from SLAC Computing) says that apparently LSF will not work with Server 2008. He will investigate. If not, we may have to rely on virtual machines.
Snow Leopard (Tom S.) (from email late last week)
I got the static version of Qt to build and the tools to compile. I retriggered the 3294 LATEST build of ScienceTools on SnowLeopard and it mostly worked. Everything compiled and most of the tests works. However, the test_optimizers test seems to have gotten lost somehow. It had run for 50 minutes fully utilizing one CPU on the Snow Leopard machine before I killed it see if the rest of the process would complete. I'm assuming it shouldn't take that long.
Announcements
GO GIANTS!
previous
|
minutes index
|
next
 |
J. Bogart, Last Modified: