Socorro: Jeff, Joe, Lindsey, George, Susan, Tak, Jim, Kumar, Sanjay
CV: Mark, Darrell, Akeem, Andy
Garching: Justo, Julian, Sandra
News / Meetings / Visitors
Reminder: No meeting next two weeks
Build, Release, Testing
Kana: Is anyone relying on OSX 10.8 for build: * Justo is using , Pam 10.8
MPI and Fortran semantics for complex multiply and divide are in the build.
Mac 10.8.3 works, but 10.8.4 does not
Nothing special at this point.
Tests against mpi enabled tarball, no mpi case, OK
Tests against 4.4 stable, slow down in image analysis method call for flux, mask images, generated by clean, possible locking problem ?
First 4.3 release bug reports, NaN problems, other items
Meeting with ALMA PWG chair
Two days building with MPI C++
MS Selection request: Feed - refactor Antenna
Having problems running mpicasa from the casa binaries in Socorro. There seems to be a problem with env variable setting in mpicasa. It fails during opal_init, OPAL_PREFIX, OMPI_MCA_mca_component_path.
Testing Justo's fixes to MPIClient performance issues.
Working on CAS-6805, importasdm should not append to online flags file.
Gave input to Dave on the implementation of msmd.summary (CAS-5053).
pipeline bandpass code cleanup
pipeline release debugging
pipeline CASA 4.4 testing / mpi enabled CASA 4.4 testing
meeting with ALMA PWG chair
pipeline results document generation design
draft of newsletter article
Worked with Kumar on velocities in VisBufferUtil
Ran plotms performance tests with old VI/VB vs. MSTransformVI/VB, then implemented AveragingTvi2 to compare its performance
Worked mainly on debugging the application wprojection in msuvbin
added some new functionality for Pam in VisBufferUtil to get velocity for vb2...this seem to work mechanically but giving different values from old vb version.
reported to Jim (vb2.getFrequencies being the culprit).
Long e-mail discussion with some CHILES member in Australia about the non need of cvel, split and mstransform for just imaging.
Justo González Villalba
Completed update of MSTransform Vis. Iter. to account for weight spectrum changes in 4.2.2 and 4.3
Fixed mpi4casa performance issues for python code executed at the client
Working on an interface class for the MPI framework that emulates parallel_go/simple_cluster
Got RHEL6 setup on new computer.
Chased after various VI/VB2 bugs; hunt still in progress.
Histogram Work with Amanda
Layout and Window Work
Casacore Google Code Compile/Install Libraries
ALMA Polarization EOC mtg, plus discussions with visitors
Added a hidden task parameter to switch off to not to raise exception for NULL MS selection for MMS
test_setjy update: Looked for replacement of 2528.ms (one of sejty unit test data sets with only 2-channel WVR spw) to one's reflect current ALMA data (->replaced with TWHya data set, updated the relevant tests) , modified all the SetJy unit tests so that it can be tested on MMS cases
on-going: fixing temporary cl file access issue for the MMS cases
Friday NAOJ Meeting
establish data reduction path of new multi-beam reciever at Nobeyama ( CAS-6987): fixed a bug in sdimaging. It should work now. Reporter is running test and investigating the results.
Improved documents and fixed a typo in sdimaging based on the tester comments on CAS-6704. The test results themselves were fine.
Draft of single dish article for CASA news letter.
Testing the change to complex mul/div rules in ASAP based on the request by Julian.
All the subticket of sd task interface changes mainly worked on CASA 4.2.2 developement cycle is closed ( CAS-5526). Thank you very much for all your excellent work!
CAS-6702, 6703: User testing finished successfully. The tickets have been resolved as done.
CAS-6997, 6998: Implementing tsdbaseline unit tests.
CAS-7001: Reading NewCalTable and related codes for implementing the new SD baseline table.
coped with a CASA pipeline (test&release) installation trouble with a load of help from Takeshi & Shinnosuke (done)
CAS-6991: Ready To Test. Generated test data and defined unit tests for sky calibration. Implementation has changed from gencal type (cb.specify) to gaincal type (cb.solve).
CAS-7266: Confirmed that Kumar's fix is working.
bug report from CSIRO ASAP user: couldn't reproduce the bug on casapy. Maybe specific to standalone ASAP?
Verified libsakura rpm installation
1 regression error and 3 unit test errors are still not fixed. Investigating these is ongoing. This is due to ScanTable has uInt column and CASA can handle this as Int but gc-cascore can handle uInt so I have to add this functionality to casa-code and gcwrap.