Core Minutes 10/12/2010

• Reprocessing: (Richard)

  • p6 ft1 repro resumes with dataclean on Tom's return. Validation apparently complete. Dataclean irf's now ready.
  • p7 ft1 repro when Markus completes definition of new diffuse class ("soon")
  • gcr files need to be remade to fill in cal calibs (can happen anytime)
  • full p6 repro with revise cal calibs (big deal, taking months and hundreds of Tbytes — needs scheduling!)
  • p7 merit, ft1 repro from p6 repro

• ScienceTools: (Jim)

  • I added the P6_V3_DATACLEAN irfs to the caldb package. Once we have the official go-ahead from LAT/FSSC software/data steering group, these irfs will be added to the caldb_release package for the FSSC to import.
  • At Jean's request, I added the capability to sub-select energy ranges on-the-fly in the python interface to the binned likelihood calculation.

• FSSC: (Eric W.) Quiet since port last week of v9r18p4

• Documentation: (Chuck) sends the following:

  With the exception of Communication Tools (which haven't been linked in yet) and the Event Viewers (which are linked only to existing documentation and can't be updated until we have an SCons version of GlastRelease), the new Instrument Analysis II section of the workbook is ready for your review. This is a post-CMT, SCons-oriented version. The pre-SCons, CMT-version will be archived and linked in to the red navbar as Archive IA I. In the process of creating this section, I've also updated a number of pages pertaining to both the CMT and SCons, as well as to both the end-user and developer worlds. These pages can be viewed from the Change Log linked in from the middle of the workbook's main "splash" page.

  Currently, I'm modifying some of the pages created for the new section, so they can be ported to the end-user section. In fact, the current end-user section may have to be archived and replaced by a post-CMT, SCons-oriented version. Among other things, this would enable us to integrate the new SCons-oriented section more closely with the Science Tools and Science Analysis sections. Creating the new integrated end-user section should go much faster than did the Instrument Analysis II project.

  You can take a pre-release peek now. As always, your comments, concerns, and questions are welcomed, preferably by the end of this week as I would like to link in the new section early next week.

  Looking slightly further out, the updates to the ScienceTools will follow immediately on the heels of updating the end-user sections.

• Pass8: (Tracy) Pass8 code, such as the new tree-based pattern recognition, is making its way into CVS. There seem to be memory issues on Linux. He is thinking of separating sim and recon steps at least for debugging purposes.

  Meanwhile Johan has been committing updates for the new Cal clustering. He's run into problems with RootIo which Tracy, working in a similar but apparently not identical environment, does not see.

• Core talk:

  • tip (Heather)There is a new tag, tip-02-15-10, which addresses a problem seen on 64-bit builds, and there is about to be a still newer one [made shortly after the meeting]

  • Windows builds (Heather) The last CMT Windows build is stuck; will try retriggering. It's not another case of missing home directory files; that has been checked. There also seems to be a problem with SCons Windows builds.

  • More Windows affairs (Richard) Our new Windows box has been moved to the Computer Building, but we've had no word since.

    New Windows systems will come with Windows 7, but there is no timetable yet for converting existing machines.

  • G4 status (Tracy) Not much progress since Francesco demonstrated in August that the new version is still not up to snuff where multiple scattering is concerned. Waiting on Francesco to have time for it.

  • BeamtestRelease (Heather) A new build with new ldf is in the works.

  • Snow Leopard (Tom S.) We still don't have a snow leopard queue. And he still can't get the RM tools to build natively there. However, there is considerable good news as well: tiger tools work on snow leopard, and he is able to make builds by hand.

  • New Gaudi (Heather) She's currently stuck on GaudiSvc but may be getting close to resolving problems. Once she does, she can proceed with the surprising number of packages which depend on it. (Joanne) There is a separate problem, as yet to be investigated, causing the CalibSvc test program to fail. It probably has to do with the new Gaudi Time class.

  • Other SCons news (Joanne) There was a bug in facilities causing the test program of flux and test programs of packages using flux to fail. She fixed this and took the opportunity to straighten up somewhat — thereby, as Heather pointed out at the meeting, causing compile failures in LATEST builds. [The problem only affected ScienceTools builds, not GlastRelease and has since been fixed.]

    Otherwise Joanne has been working on code to generate rootmap files for SCons builds. A draft is written but has not been tried.

previous

minutes index

next

J. Bogart, Last Modified: