Monday Morning Meeting 21 September 2015

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

Attendance

* Socorro: * Socorro: Jeff, Susan, Lindsey, Pam, Sanjay, Urvashi, George, Martin, Tak, Jim * CV: * CV: Darrell, Ville, Akeem, Dave, Mark * Garching: * Garching: Julian, Sandra, Justo * SCO:

News / Meetings / Visitors

  • Chris Hales
  • ALMA Single Dish Acceptance recommended
  • Strategic Staffing Summit
  • Eduardo Closeout
  • Projects

Build, Release, Testing

  • tClean issue: Issue with find convolution function in mosaic FT
  • Blocker issue in plotMS
  • Justo:
    • Handling of ephemeris file in msTransform - Go
    • Parameter names changes - Go
    • Lindsey to send Sandra UID of example where ephemeris is used for flux calibrator.

Architecture

* Technical Forums: * Testing in in place on trunk, Google Test requires more thought.

HPC

  • Julian and Justo fixed an ineficiency in flagdata summary mode (actually in VI/VB2), which benefits the SD pipeline; what took 2 minutes to execute now takes 10 seconds.
  • Julian also suggested a patch to the SD pipeline that improves pieces of the Python code..
* Have been running performance tests using the alma-m100 regression for the next CASA newsletter. The tests use one of the latest casa-pre-release tarballs. The speedup is about 1.4x faster in parallel. This may improve further once tclean is used in the script. * Have been running performance tests using the alma-m100 regression for the next CASA newsletter. The tests use one of the latest casa-pre-release tarballs. The speedup about 1.35x faster in parallel. This may improve further once tclean is used in the script.

Development

Pipeline

  • C3R3 release preparations
    • Data import: fixed HISTORY scoring to account for changes in CASA importasdm HISTORY logging
    • Flagging: fixedto ALMA deterministic flagging statistics (QA0) reporting and scoring, fixed low gain flagging missing caltable handling
    • Imaging: sorted out model column selection and uvsub handling problems, turned on full cube imaging, followed up on PSF creation errors and memory management isses with the imaging team
    • Infrastructure: investigated boolean / string conversion bug in pipeline script generator code
    • Data export: changes to support new image products, especially multi-term images
    • Continuum region finding: integrated algorithm improvements from Todd into pipeline
  • JIRA bug report followups, Tsys flagging, web log links

AOB

Developer Reports

Monday Meeting
  • Sanjay Bhatnagar
  • Sandra Castro
    • HPC work as described above.
    • Have noticed that task uvcontsub calls ms.split() internally and we need to decide if we want a similar method in the mstransformer tool now that split will be deprecated.
  • Lindsey Davis
    • Model column / uvsub testing with Dirk M, email exchanges with Dirk M, Urvashi
    • Discussion / emails on tclean memory management with imaging team and Dirk M.
    • Fixed pipeline HISTORY scoring issue
    • Tracked down pipeline script generator bug
    • General testing
  • Pam Ford
    • CAS-7831 blocker - asked for more information than "it's not fixed"
    • CAS-2091 RFI (pm tool resets options when change data selection) - old ticket that I think is no longer an issue
    • CAS-7875 (glitchy model column when time-averaged) - issue was actually the averaged uvw values used for uvdist and uvrange, reassigned to Jim
    • CAS-7851 RTT (restart casaplotms process when it crashes without restarting casapy)
    • CAS-7242 Resolved (do not show axis when "Show X/Y Axis" checkboxes unchecked)
    • CAS-7878 Resolved - added Dirk's fix to release-4_5 branch
  • Kumar Golap
  • Justo Gonz├ílez Villalba
    • Attended ERIS interferometry school
    • Refactor DDI/SPW/Pol methods of mstransform to avoid inconsistencies reported in CAS-7814/CAS-7841
    • Fixed sign issue reported in OTF phase shifting applied via plotms
    • Updated mstransform section in CASA cookbook
    • Fixed mstransform problem handling ephemerides files
    • Implemented a generic weight propagation method in mstransform applicable to all linear combination kernels
  • Jim Jacobs
  • Jeff Kern * Strategic Staffing Summit * Eduardo Closeout
  • Susan Loveland
    • Attended Very Long process management meeting
    • CARTA
      • Worked on support for multiple axes
      • Pan, image crash, and deployment problem bug fixes
  • David Mehringer
    • 4 days in office, 1 day sick.
    • Reported imager_cmpt.cc warnings to casa-staff. ms_cmpt.cc also has several compiler warnings.
    • Reviewed some of Martin's statistics code mods.
    • [casacore] Remove extraneous call to sqrt() (#233: reviewed Martin's fix at Julian's request.
    • (CAS-7679) fitting in non-optical mode does not work when restfreq is not set: RTT, added support to specfit-related code for images which have restfreq=0.
    • (CAS-5104) LCSlicer::toSlice appears to have a defect, Resolved, Not a bug. Due to a misunderstanding I had about an LCSlicer constructor.
    • (CAS-5597) CRTF doco improvements: Changed to input required as a way to note that I cannot proceed on this ticket until work on CAS-7704 has been resolved.
    • Attended yet another marathon presentation given by Eduardo. Had to leave at the 95 minute mark.
    • (CAS-5893) add the ability to write a spectral profile to an ASCII file in the ia tool: RTT. Completed implementation of requested functionality.
    • (CAS-6112) Specfit output: Input required. Queried requester if functionality added as per CAS-5893 would be sufficient to resolve this ticket.
    • After prodding CIS, received USB hub for my new mac that I had requested a week prior. Hopefully this week I will have time to install CASA.
    • (CAS-7778) msmd should have a method to retrieve the effective bandwidth or the effective resolution of the channels rather than only their formal bandwith: RTT, implementation complete.
    • (CAS-3366) listfile parameter for imstat: RTT, modified output as requested.
    • (CAS-5996) immath unit- (and beam-) awareness: Began implementing general parser for LEL expressions in order to extract image names using a lot of regex and string manipulations.
    • (CAS-7637) msmd should have methods to retrieve REST_FREQUENCY and TRANSITION frm the SOURCE table: RTT, completed implementation.
  • George Moellenbrock
    • ALMA Polarization testing/discussions w/ Crystal --- fixed (in release) a problem with feed pa discovery triggered by a spw name convention change at ALMA
    • Cookbook Cal chapter edits for 4.5
    • Various other user support/testing, most notably a query re spurious solutions from blcal from Robert Laing on a 500000:1 imaging project.
    • Adapting to RH6 on my dev machine....
  • Dirk Petry
  • Martin Pokorny
    • Submitted pull request to casacore for changes needed for visstat2; later added unit tests for changes, as well.
    • visstat2 C++ code largely complete; a few MS columns not yet supported; needs work on interface and integration with task layer.
  • Urvashi Rao
    • Implemented cube chunking for large cubes via reference subimages.
    • Tested multi-field across chan ranges and isolated some issues.
    • Finished inline docs and iterating with CB on edits
    • Emails about continuum subtraction and model columns
  • Darrell Schiebel
  • Ville Suoranta
  • Julian Taylor
    • tested new mpi dependencies of casa binary rpms (the rpms with stuff in /usr/bin)
    • removed unnecessary getFrequency calls from flagdata, factor 10 performance improvement for some summary mode cases
    • improved performance of visbuffer::getFrequency by removing unnecessary Unit object creations
    • reduced cost of Unit object creation in casacore
    • started looking at some timeout issues with sdpipeline, found a couple simple but significant performance improvements
  • Takahiro Tsutsumi
    • Added a unit test for setjy for the data with ephemeris table
    • Updated tclean inline documentation for cube imaging sub-parameters
    • Parallel cube tclean testing, found an issue (appeared that it was not going through request iterations in minor cycle).
Friday NAOJ Meeting
  • Kanako Sugimoto
    • A summary presentation of CASA 4.5 development cycle. Useful feedback from team members. It turned out short testing period by quick release process is making Erik's life hard.
    • Definining CASA 4.6 targets. Started discussion of plotter developments.
    • Discussion of the status of Solar fast scanning observations with Stephen White and Kazumasa Iwai.
    • 21 Mon - 23 Wed: National holidays of Japan. You may not get active responses from NAOJ developers.
  • Wataru Kawasaki
    • made refactoring of SD baselining codes both for C++ and Python (tsdbaseline task code) sides since they have been quite lengthy.
  • Masaya Kuniyoshi
    • added intent to visstat2.
    • refactoring unit tests of tsdbaseline.
  • Takeshi Nakazato
    • Discussion with Kana and Erik about CAS-7452 and CAS-7451
      • As for CAS-7452, we agreed that we implement Scantable to MS filler first, and then NRO data filler follows
      • Anyway filler should be designed so that it is flexible enough to be able to implement NRO data filler easily
  • Renaud Miel
    • CAS-7339
      • Reviewed task_sdcal2.py and task_sdcal.py code
      • Fixed syntax error bug in sdutil.py
      • Identified out-of-date documentation issue regarding calibrater.setsolve method, calibration 'type' parameter

-- KanaSugimoto - 2015-09-17-- KanaSugimoto - 2015-09-17
Topic revision: r22 - 2015-09-21, 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