Core Minutes 11/8/2011ScienceTools: (Jim) There is a new Likelihood, implementing energy dispersion in gtmodel.
Steve Fegan looked into the diffuse response integrals to see if they played some role in the differences seen between the binned and unbinned analyses. He was to give a presentation on this at the Monday C&A meeting, but we ran out of time. He may discuss it at the next meeting or later this week.
FSSC: (Eric W.) Just wrapped up an overhal of the web interface to the database. Next on the horizon: the long-delayed tip updates. He's been in touch with James Peachy. It will be a substantial amount of work — to be measured in weeks — but he expects to be able to focus on it. This will allow us to advance to a new version of cfitsio; the one we're currently using is quite old.
Pass7: to reprocess or not to reprocess: (Leon) It's still up in the air, expected to come to a head in the next week or two.
Systests for SCons: (Leon) Liz has taken the bait. She'll help us get going, if need be with a semi-manual approach to start.
GR/Pass8 (Tracy) It's fine, but not much to report since last week. It will be another week or two before the next tag, which will implement the new sequencing of recon steps.
(Leon) gave Bill's Pass8 talk at the IFC meeting since Bill was unable to give it himself. Audience reception appeared favorable. Several preceding talks set the stage by referring to it.
Demise of redhat 4 John Bartelt has contacted us about moving glastlnx14 and glastlnx15 (our RHEL4 build boxes) to a new queue, bld4q. This is in preparation for SLAC's move to drop RHEL4 support as of December 9th. These boxes are used by both CMT and SCons RM for the RHEL4 builds. Kim and Tom are both aware of the move, due to happen by November 15th.
GR externals and redhat 5 (Heather) She needs to finish up making a Red Hat 5 Gaudi build. There is nothing especially daunting about it; it just takes time. (Joanne) made a new TMineExt tag, TMineExt-03-02-05, so that the corresponding external can be properly built from a production tag (has not been the case previously). The new external is installed for redhat5 32-bit and 64-bit as well as redhat4 32-bit. Ditto for obf B3-0-0.
ST and redhat 6 RM (Tom S.) was able to build all the RM programs on rhel6 and has verified they work correctly. Initial attempts at building ST failed because of externals problems. (Heather) Kim redid ROOT and python to get them into the proper form. [Late-breaking news: there was one more problem with the python external, now fixed. LATEST-1-3634 mostly built, but there appears to be a problem with f2c. Heather is hot on the trail.]
SCons Windows RM (Tom S.) may have found and fixed the problem. He noticed that builds would succeed while he was logged on via remote desktop, as if some token or key was available then but not necessarily otherwise. The missing bit appears to be an enivronment variable, APPDATA. He added it in on glastlnx20, which launches RM jobs. Since then (Friday) jobs have been running successfully.
obf on Windows (Joanne) All that remains is to test, but that requires a version of GR and externals which will run with VS 2008/vc 9.0. Several heretofore missing externals have been retrieved or built (G4 version 9.4, TMineExt, zlib). Now the problem is with GR itself. Many updates required for the new gaudi (the only one we have for VS 2008) have not been merged into HEAD; they only exist on a branch of GR v18 vintage. (Heather) will deal with it one way or another: either bite the bullet and merge changes into the main branch (has been difficult to find a time for this because of Pass8 activity), or make a newer new-Gaudi branch.
glast-ts and related matters (Toby) was wondering how much longer glast-ts (which runs Windows Server 2003) will be needed; that is, when all work can be switched to Fermi-ts, which runs Windows Server 2008 with VS 2008. (Heather) It hinges on dropping CMT. This is about to happen on Linux because of the end of redhat 4. It will happen on Windows as well as soon as we can build a usable GR with the newer compiler (supported only by SCons). We're getting close; see previous item.
(Heather) was wondering aloud about the need to test our GlastRelease and ScienceTools builds on Windows 7 (or Windows Server 2008) to see that they work in that environment. (Toby) Windows 7 just adds nice user interface features so most likely will not cause compatibility problems. (Heather) Should we be supporting 64 bit builds on Windows? (Toby) sees no necessity since Windows is used for development only [and no one else chimed in, so likely we will continue to do 32 bit builds for Windows until there comes a push to move to 64 bit].
Announcements
Richard, Tony J. and Rob are away this week. Anders is holding down the ISOC fort.
SCons meeting Wednesday at 10:30
|
|
minutes index
|
next
|