Monday Morning Meeting Dec 15, 2014

  • DIAL-IN NUMBERS & PASSCODES:
  • USA Toll Free Number: 866-901-8266
  • USA Toll Number: +1-203-566-3863
  • PARTICIPANT PASSCODE: 725760

Attendance

  • Socorro: Jeff, Joe, Lindsey, Pam, Juergen, Kumar, Tak, Susan, George, Jim
  • CV: Akeem, Andy, Darrell, Dave, Mark
  • Garching: Justo, Dirk, Julian
  • SCO:

News / Meetings / Visitors

  • Joe Depa in Socorro
  • Last Meeting of the Year
  • Colloquium this week:

Build, Release, Testing

  • CASA Data
    • Update Scripts are running, every Sunday. Runs on EL5, so rapidly becoming out of date.
    • Last updated 2 months ago.
    • Need to update the scripts.

  • 4.3 Release process
    • CAS-7176: Good to go on Linux
    • Updating ALMA ATM Library
    • OSX - Working on the Lattice Issues
      • Adding standard tools to the Mac Test System
      • Old VLA Filler seems to be failing in NGC-2403 test.

Architecture

  • Technical Forums:

Development

  • Systemic Velocity: IRAM want's systemic velocity from source table in the images.
    • Joe to work with Dirk, Kumar, Juergen to do a requirements capture.

Pipeline

  • CASA testing
    • applycal slow down has not re-occurred
    • plotbandpass issues fixed
    • importasdm XML reading issue fixed
  • pipeline testing
    • plotting speedup factor of 2~3 in total execution time
    • calibrator catalog flux retrieval integration

AOB

Developer Reports

Monday Meeting
  • Sanjay Bhatnagar
  • Sandra Castro
    • I'm attending a full-day workshop today and cannot attend the Monday meeting.
    • Created a ticket to have functional tests for applycal. We need reference tests on MSs in order to test applycal on Multi-MSs. CAS-7199.
    • Resuming my MMS tests on the pipeline. The seg fault on plotms is "gone". I build the pipeline with the release-4_3 branch. I'm analyzing setjy to see how it can be improved to run on MMSs.
    • Merged fixes to the consolidation of some sub-tables from trunk to release 4.3 branch. These changes fix problems when the WVR spw is alone in a sub-MS.
  • Lindsey Davis
    • improvements to low SNR bandpass calibration - antenna flagging support
    • applycal testing
    • general release testing
  • Pam Ford
    • Worked on implementing VI/VB2 in plotms with various data selections (no averaging yet) as a first step.
  • Kumar Golap
    • Worked mainly on understanding some ALMA data mosaic slowness
      • Wrote code to cache the pointing for all antenna at all integration time
    • Fixed a crash with WProjection with large cubes that Emmanuel found
  • Justo González Villalba
    • Fixed cluster issues reported by Jenkins for MACOX 10.8: It was necessary to update foolscap and add OpenSSL
    • Investigated flagdata/flagcmd issues reported by Jenkis for all OS:
      • It was a problem due to the recent mods in VI/VB tile caching, I've tested a mod provided by Jim which fixes the problem
    • Worked on alma-cycle 2 parallel regression case using the maximum number of sub-MS possible: 428 sub-MSs
      • Found issues in wvrgcal which crashes after seg. fault. when processing a MMS as a MS. I've reported this to Dirk (CAS-7200)
      • Found performance issues in flagdata running in list mode (11190 commands): It seems that even if the MSSElection parsing time is not too large for this case (5mins) it is a big penalty to pay this time for each sub-MS when the number of sub-MS is quite large. I've created a ticket for it (CAS-7201)
    • Tested Julian's OpenMPI /mpi4py rmps for RHEL 5.5
      • Found an issue with CMAKE that is not setting the library paths properly and it is necessary to set LD_LIBRARY_PATH manually
  • Jim Jacobs
  • Jeff Kern
    • Wrote response to CASA Users Committee.
  • Susan Loveland
    • CARTA make problems (help from Jim)
    • CARTA colormap & histogram GUI.
    • CARTA crash problems
  • David Mehringer
    • Committed first phase of stats framework which includes ClassicalStatistics class. Modified LatticeStatistics to use new stats framework which provides significant performance improvement over previous implementation.
  • George Moellenbrock
    • CAS-7176: continued testing of applycal; added fixes for channel selection and the uvcontsub context. No new reports of slowness since Jim's casa tile cache fix.
    • Continued investigations of I/O patterns underlying applycal (prompted by CAS-7176), most notably concerning the difference in physical data sort in DATA and CORRECTED_DATA.
  • Dirk Petry
    • 4.4 Development: System Velocity, related changes to import/export UVFITS, extension of import FITS IDI
  • Urvashi Rao
  • Darrell Schiebel
  • Julian Taylor
  • Takahiro Tsutsumi
    • Resumed work on automask (worked on framework, extended a test version simple autoboxing code to a cube)
    • Wrote a python function to extract information from the fluxscale returned dictionary and to format them to write to HISTORY subtable (CAS-2152)
    • Commented on CAS-6991 regarding to the current caltable format
  • Ger van Diepen
    • Created nov14 branch of Googlecode casacore
      • added casacore/ to #include paths to get standard -I paths
      • changed namespace casa to casacore
      • split Tables into Tables, DataMan and TaQL
      • Backward compatibility is available for, say, one year
      • The nov14 branch will be the basis for the unification
      • Casacore will move to github; Malte is preparing the move
      • Helped Shinnosuke in building CASA against this branch. When building CASA I see very many warnings of which some look serious; e.g. uninitialized variable or use parentheses in mix of || and &&. They should be taken seriously; in Casacore the || && parentheses warnings lead to 4 bug fixes, because some think that || and && have equal precedence.
    • All NRAO changes in casacore till 9-Dec-14 are merged in
    • Started working on article for special data model edition of A&C
Friday NAOJ Meeting
  • Kanako Sugimoto
    • work on user test feed back: 6564, CAS-6467, CAS-6704, CAS-6587, CAS-5139 (one more to go: CAS-6587)
    • generation and managements of project-singledish-ms branch
    • CAS-6989: sdms tool - need to create a framework to generate output ms. looking into mstransform codes.
    • discussion of information necessary to process fast scanning data: put comments to tickets: ICT-4099, ICT-2437 (observation mode)
    • PICK discussion items: cal table definition, osx support, casacore namespace, seg fault of sdcal2
  • Wataru Kawasaki
    • CAS-6997: implemented an early version of sdms.subtract_baseline()
    • CAS-6997: compiled a parameter list for tsdbaseline, a new SD task for baselining, then implementing its xml task interface (yet to be finished)
    • modified sdms.get_spectrum_from_cube() so that a spetrum value can be obtained as an aligned array (SakuraAlignedArray)
    • installed RHEL6.6 in the two desktop computers for CASA development at NAOJ, which will be used by Masaya and me
  • Masaya Kuniyoshi
    • CAS-6992 [MS transition: Define a task to generate Tsys calibration table]
      • discussion of the tsdcal interface: the same as the sdcal2 task, which can handle asap files and generate Tsys tables
      • discussion of the tsdcal interface: the same as the sdcal2 task
      • strategy: taking advantage of the gencal task loading MSs and producing Tsys tables
      • learning how to import other tasks, etc.
      • started coding
  • Takeshi Nakazato
    • bug fix on sub_baseline
    • investigation on segmentation fault of sdcal2 (CAS-7193)
    • preparation for sky caltable definition (CAS-6991): wait for reply from George/Jeff about caltable definition document compatible with current implementation
    • formulation of weight update for calibrated data based on Bunyo's document, which is attached to the ticket (CAS-6994)
    • attempting build singledish-ms-new branch with Sakura on Mac OSX (10.8)
  • Shinnosuke Kawakami
    • CAS-6929: Created branch from casa-code trunk (r31805) to commit local change of the code-set.

-- JeffKern - 2014-12-08-- JeffKern - 2014-12-08
Topic revision: r20 - 2014-12-15, JeffKern
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback