Monday Morning Meeting DATE

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

Attendance

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

News / Meetings / Visitors

  • DMS Leads Face to Face
  • Huib van Langevelde Visit

Build, Release, Testing

  • CASA 4.5 Release:
    • CAS-7831: Issues in frequency VLSR
    • CAS-7914: may be same thing as above

Architecture

  • CasaCore Namespace: CasaCore changed over to using its own namespace, casacore, rather than the shared one, casa. The change was done months ago but a hack (#define casacore casa) was provided to keep from breaking CasaCore users such as CASA. Ger wants so remove the hack so that CasaCore will truly reside in its own namespace. The CASA projects code, gcwrap and asap will require modification to allow them to build with the CasaCore namespace. A script can be constructed which will effectively add a "using namespace casacore" directive to all of our .cc files. Initially, the change will have little effect, but as CasaCore evolves, the independence of the separate namespaces will remove the chances for collisions with CASA and other users from seemingly backwards-compatible modifications to CasaCore (e.g., adding a new class to CasaCore such as Calibrater, etc.).

  • Redundant Typedefs As part of the CasaCore namespace conversion, the use of CASA's own names for standard types (e.g., Bool, Int, etc.) has come into question. Does anyone know of any problem with replacing those types with the standard ones? String is a special case, since it bundles together a lot of functionality relative to std::string. Others such as Complex (= std::complex) are probably handy since they shorten things.

  • Architecture Page link The CASA Architecture Page has links to info on warnings and the above link to testing.

HPC

  • Sent to Jeff a text for the CASA newsletter.
  • Working on the close-out items for the HPC sub-project and the 4.6 development items.
  • Re-created the MMS datasets for all functional tests that run in the ESO jenkins. Improved the convertToMMS() class inside partitionhelper.py that is used to create these tests.
  • Discussions about the performance tests using the alma-m100-hpc script.

Development

Pipeline

  • Chris Hales began work with the pipeline on SRDP
  • C3R2 trunk prerelease made available for ALMA calibration and flagging pipeline validation by the PWG * ongoing testing, bug fixes, display fixes
  • CASA 4.5 / pipeline C3R3 tclean testing * infinite loop bug fixed, PSF creation bug fixed, common primary beam technical issues fixed * new masking issues ?
  • EVLA pipeline development goals identified * low SNR calibrator heuristics, polarization calibration, syspower issues
  • ALMA pipeline issues * end to end coordinates handling, what are J2000 coordinates in CASA * session handling

AOB

Developer Reports

Monday Meeting
  • Sanjay Bhatnagar
  • Sandra Castro
    • I cannot attend the telecon today!
    • HPC work as described above.
    • Tested Sanjay’s fix to MS selection in casacore with respect to supporting spw names.
    • Tested Lindsey's ASDM with ephemerides to create an MMS. Did not see any issues.
  • Lindsey Davis
    • tested ALMA imaging recipe
    • monitored pipeline C3R3 JIRA ticket
    • email on ALMA end to end coordinates handling issues
    • helped Chris Hales install and test the pipeline on his local machine
    • drafted pipeline article for next newsletter
  • Pam Ford
    • CAS-7831 blocker (virtual model) - investigated and asked Kumar to look at virtual model generation; problem is in VB2 getFrequencies()
    • CAS-3125 RTT (unzoom bug when iterating) - fixed zoom stack base plot
    • CAS-4408 Resolved, Cannot Reproduce (old ticket about DEBUGGING messages)
    • CAS-3091 Resolved (duplicate ticket for OSX display issues)
    • Created developer build on machine in CV for debugging with access to CV lustre datasets
    • Investigated sdplot plotting requirements from Kana to implement in plotms
  • Kumar Golap
  • Justo González Villalba
    • Ported fixes to 4.5 branch
    • Fixed a couple of problems related with LogFiltering in the HPC context
    • Fixed a common performance problem in mstransform related with POINTING sub-table handling
    • Investigated phase shifting differences between mstransform approach and fixvis
    • Contributed to HPC newsletter, M100 regression performance analysis, etc
  • Jim Jacobs
  • Jeff Kern
  • Susan Loveland
    • Viewer: CAS-7902: Spectral viewer shows wrong result for small elliptical regions (ready to test)
    • CARTA:
      • Miscellaneous enchancement & bug fix requests in PR 79 & 83
      • Axis work
  • David Mehringer
    • Fixed memory leak and open table issue in ImageUtilities::openImage()
    • (CAS-5407) immath should reorder degenerate axes: Assigned for input because implicitly reordering axes of images seems heavy handed, and judging from past CSSC recommendations on immath, seems like is more than that is likely wanted by them.
    • (CAS-7087) Parent ticket for UVFits Bug Fixes for 4.5 Cycle: Assigned to Mark for management as I am currently not assigned to any of the remaining open tickets.
    • Reimplemented image name extraction from LEL expression method to avoid using C++11 regex functionality which fails on RHEL5 because that version of gcc doesn't fully support C++11 regex.
    • Communicated with Tak re: ImageRegrid.
    • (CAS-5893) add the ability to write a spectral profile to an ASCII file in the ia tool: RTT. Completed implementation.
    • Added methods to ImageProxy to return ImageInfo and CoordinateSystem objects.
    • Worked with Ville (thanks Ville!) to install development environment on my new Macbook. I can build casa from scratch in under 20 minutes on it, which I believe is a new personal record.
    • (CAS-7913) Fix MSMetaData::getScansForTimes on OSX: Submitted pull request to fix segfault on Mac. Later was told a comment was required in the code to clarify what I did, but by that time I had moved on and was immersed in something completely different. So, if the current state of the code isn't good enough to be merged, it will have to wait until I'm once again back in that file as switching context to add a single comment simply is too expensive from a productivity point of view.
    • (CAS-5688) Performance improvement for ia.modify(): Much work trying to understand how this is implemented. For the use case given, there are of order 10 billion unnecessary sqrt and asin calculations being done, so I have to work out a much more efficient implementation.
  • George Moellenbrock
    • More ALMA Poln support
    • NRAO "Space Morning" at Parkview Elementary School
    • New cookbook edits
    • Got my RH6 build working
    • Began calibration code survey, for testing, deprecation, etc.
  • Dirk Petry
  • Martin Pokorny
  • Urvashi Rao
    • Finished inline docs for tclean (and responses to feedback from CB)
    • Fixed bug in iteration control
    • Worked on chapters for synthesis imaging book
  • Darrell Schiebel
  • Ville Suoranta
  • Julian Taylor
  • Takahiro Tsutsumi
    • Worked on the issue that an input cube mask not correctly resampled in some cases (CAS-7899). Casacore's ImageRegrid using it directly seems to have the issue. By replacing it with ImageRegridder solved the issue in some cases, but not all. Work continue to make it work generally.
    • Fixed a bug in uvcontsub's excludechans (CAS-7888)
    • Fixed doxygen.cfg so the c++ documentation can be built properly
    • Still working on: parallel cube tclean
Friday NAOJ Meeting
  • Kanako Sugimoto
  • Wataru Kawasaki
  • Masaya Kuniyoshi
  • Takeshi Nakazato
  • Renaud Miel
Topic revision: r13 - 2015-09-28, 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