Monday Morning Meeting 04/30/18

  • DIAL-IN NUMBERS & PASSCODES:
  • IP: 192.33.117.12##8110
  • Phone: (434) 817-6524

Attendance

  • Socorro:
  • CV:
  • Garching:
  • SCO:

News / Meetings / Visitors

  • Pipeline F2F: Tokyo, 5/7 - 5/11
  • NRAO Users Committee Meeting: Socorro, 5/14 - 5/16
  • CASA Newsletter, 5/9

Build, Release

  • Atlassian maintenance starts at 3PM EDT today.

Development

  • 5.3 issues
  • Obs mode 7 in Chile?
  • CASA Software Design Document (SDD):
    • Created template in Libre Office, saved in .fodt format with SVG graphics from draw.io
    • Attempted git check in and test merging of disparate changes - failed
    • This week: try Markdown format
    • Can distribute fodt template in the meantime if people want it

Pipeline

  • Pipeline F2F meeting preparations
  • CASA 5.3 issues
    • mfs auto-masking imaging repeatability issue

HPC

  • Differences serial/parallel
    • CAS-11322- "Blockers for parallel use of ALMA pipeline"
      • Differences in beam are >1% only for ~3% of the test images (>600), although there is a handful of outliers.
    • CAS-11321- "source fit warning seen in serial, not seen in parallel pipeline runs during 5.2 testing"
      • The "Check source fit" warnings are present in the logs from serial and parallel runs, and they match.
      • They're also highlighted in the "By task" summary page of the weblogs, both serial and parallel, but they don't show up in the details page for task 20 in the weblog from the parallel run. Same with more recent prereleases. See ticket for details.
      • Looks like the parser that produces these messages in the weblogs is getting confused in parallel mode (log lines have different prefix)?
      • Same warnings, but numerical differences throughout CASA versions (5.1, 5.2, 5.3).
  • Automated testing
    • I think parallel-tclean should be tested/validated outside the pipeline. I discussed with Urvashi who agreed with me that the serial tests of tclean (test_refimager) could be slightly modified to run in parallel too. It is a simple matter of detecting if an mpi cluster is available and setting the parameter parallel=True to all tests. I'm trying this approach and will share the logs with the imaging team. All 159 existing tests should in principle handle the situation when parallel is set to True in the interface.
  • 5.3 Items
    • Committed fix to enable task_initweights to run in parallel (CAS-11262).

Users

  • Newsletter articles ready: CAS-11159. Feedback welcome until end of the week. Please send me user-facing CASA events for the "Calendar".
  • CASA Docs: please review "Release Notes" and "Known Issues" on development server.

Verification Testing

Validation Testing

Architecture

AOB

Developer Reports

Monday Meeting
  • Sanjay Bhatnagar
  • Sandra Castro
  • Lindsey Davis
    • Investigated mfs auto-masking repeatability issues with Dirk Muders, confirmed there are issue(s)
    • Pipeline F2F preparations
    • Obsmode 7 planning discussions
  • Bjorn Emonts
  • Pam Ford
  • Enrique Garcia
  • Bob Garwood
  • Kumar Golap
  • David Mehringer
  • George Moellenbrock
  • Dirk Petry
  • Martin Pokorny
  • Federico M Pouzols
    • CAS-10664 - "Object name not in header for most images from tclean with mtmfs enabled"
    • Testing/following:
    • CAS-10392 = "Python MemoryError: tclean respects memory limits but pipeline code is more memory hungry?"
    • CAS-11322 - "Blockers for parallel use of ALMA pipeline"
    • CAS-11308 - "Error in task hifa_imageprecheck: invalid literal for int() with base 10: '5,7,9,11'"
    • CAS-11321 - "source fit warning seen in serial, not seen in parallel pipeline runs during 5.2 testing"
    • CAS-11348 - "pipeline.hif.tasks.tclean: problem with extra quotes - Field Expression: No match found for name "J1751+0939" (near char. 0 in string ""J1751+0939"")"
  • Urvashi Rao
  • Darrell Schiebel
  • Kanako Sugimoto
  • Ville Suoranta
  • Takahiro Tsutsumi
Friday NAOJ Meeting
  • Wataru Kawasaki
  • Takeshi Nakazato
    • CAS-11333sdimaging bug: wrong position reference for frequency conversion [Ready To Validate]
      • could cause channel shift for high velocity resolution data. clean is potentially affected, too.
    • testing CAS-11306(plotms error)
      • test package worked fine. prerelease package worked fine, too.
    • investigation on plotms atm transmission overlay (probably I create a ticket)
      • altitude hardcoded (ALMA)
      • unit handling for pressure is missing (while temperature does care about the unit)
    • CAS-11338 listed up single dish tasks that require update
  • Renaud Miel

-- TakeshiNakazato - 2018-04-25
Topic revision: r8 - 2018-04-30, BjornEmonts
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