Core Minutes 11/06/2007

• GR releases: (Heather) HEAD has ACD calibration changes backed out; CAL updates from Zach, including mods to Trigger, are in, along with mods for use of gr_app to many packages and some tags from Martin, rationalizing use of GEM and GLT. Part of this involved changes to the ILivetimeSvc interface, incidentally causing problems with Interleave.
  • Trigger (Martin) The idea is to include GEM information in MC. Trigger will now
    1. apply window mask
    2. prescale by engine
    3. kill events as appropriate in consideration of livetime

    He had thought the livetime counters in ILiveTimeSvc were suitable and was not aware of their different use by Interleave. (Toby) Since it's a service, not an algorithm, another function could be added to the interface so that both uses could be supported. (Martin) will proceed in this direction. He estimates it's only a few hours' work. It should also take into account the different deadtime for 4-range readout. Subsequent email from Anders adds more detail:

    Normal readout: 529 ticks
    4-range, suppressed readout: 1308 ticks
    A tick is 50ns.

  • CAL (Zach) The initial goal was to more realistically simulate trigger decision with regard to CAL input; for example to output readout mode rather than having it be fixed at initialization time through job options. But the current code organization wouldn't support this; it would lead to a circular dependency, among other problems. Solution was comprehensive reorganization so that simulation behaves much more like the instrument. Break out the new CalSignalTool to do digitization for diodes. (This is not "official" output in that it is not stored in the TDS nor written to a file). This information is used both by TriggerAlg and also by CalDigiAlg when it generates the CalDigi collection. The new scheme has minimal impact on job options – as minimal as it can be, that is. Old job options will work as they used to, but to make use of the new functionality one must change the order of execution. (Toby) Changing basicOptions should do the trick for most applications. (Heather) This should accompany Zach's work in the next GR tag. See also Zach's admirably clear and complete exposition, including diagrams which may be found in the CalXtalResponse doc directory. To finish off the project, Zach is about to commit code to generate 4-range readout, needed to simulate GCR calibration input.
• Big run: (Richard) See the blog for late-breaking news.

  We'll have to freeze GR – that is, work from a branch – for something like 6 weeks during the generation of data. (Heather) doesn't seen any intractable difficulties; we've done this before. We know how to trigger a build of a tag along a branch. Should confirm that we can run Sys Tests as needed on tags or HEAD builds.

  (Toby) put in some new NTuple variables at Bill's request, to be used to calculate neutral energy and hopefully improve PSF. No feedback yet. (Tracy) The ntuple is approaching recon data in size: it's about 25%. Does it make sense to keep throwing more into the tuple rather than using recon directly?

  (Heather) CEL should be in within a couple weeks.

  (Richard) The somewhat alarming CAL and SVAC size estimates from last week were inadvertantly inflated by about a factor of 5, so there is no pressing need to eliminate them. David will be enhancing the Skimmer to handle them.

• Navid's world   
  • extlibs Will be adding another libary, 4SuiteXml, to the python distribution. There is a new extlib ape which will be replacing pil.
  • SCons status update Created a swig wrapper for facilities to make functions available to python. It worked as intended on Linux. On Windows SwigPolicy and RootcintPolicy have different behavior: they modify project files. This can cause collisions in a package like commonRootData which uses facilities and RootcintPolicy when one tries to build in MRvcmt just by clicking the build button rather than doing the build in the Windows IDE. Navid managed to work around it by putting the use of SwigPolicy in the private section and the use of swig itself in the public section of the facilities requirements file.

    Jim Chiang has volunteered to take care of hardcoded references to environment variables in python code. A similar job for XML will come next. Once ScienceTools is fully functional, including testing on Windows and Mac, Navid will start in seriously on GlastRelease. Meanwhile he's beginning work on the revamped RM, e.g. designing the new MySQL tables which will be needed. See also the Todo list.

• AOB: (Richard) Thanks to Anders and Zach, respectively, we have updated documentation for the SVAC tuple and CAL tuple.

previous

minutes index

next

J. Bogart, Last Modified: 01-Jun-2010 15:46:59 -0700