ALMA computing meeting last week. Debrief from Jeff when he gets back.
Build, Release, Testing
4.3 RHEL pre-release packages under construction.
Still one test failing, but infrastructure is in place.
4.3 OS X Pre-release / stable package
Darrell to start today.
EL5/EL6 using 4.2.22 stable (4.3 pre-release)
Test machines in CV running RPMs, so they'll have to use the latest stable. Eventually shift to pre-release when available.
Deployed new RPMs on casa-dev nodes and casa team workstations in the DSOC.
Completing move to new package names, version numbers, launchers.
Notice to user testers tomorrow.
Status of publishing changes to users (AOC, CV, ESO, NAOJ)?
Exchanges with Kana helped clarify documentation.
Need to contact ESO, Chile, AOC, etc.
Juergen and Mark to work on NM Sci announcement.
Before putting on to lustre in CV, Darrell needs to put in links to look for local tars unpacked in home directory.
JIRA workflow - Rob to add a way to reopen tickets.
Andy is close to finding a candidate for the testing group. Committee close to a decision.
Need to put links to regression test results on the casa-wiki.
CASACore path forward - Jeff & Jim - (TBD)
B&T Requirements - Mark
B&T System Redesign - Darrell
TClean - Urvashi, Sanjay
Imaging Pipeline -
Structure of flagging classes and flagging algorithms - Sandra or Urvashi?
Vi/VB2 implementation (already presented VI/VB, but we could focus on remaining work to complete migration) - Jim ?
CARTA Architecture (when ready for discussion)
ADMIT Project - Jeff
MSTransform heuristics & MMS - Sandra - Nov 13?
Still trying to get a clean build as a prerequisite.
After a clean build, can consider the mechanics of repository integration.
Ger is working on merging changes over.
Statistics Framework requirements/Implementation
Status of discussion regarding large data sets?
Jim to follow up.
Next HPC meeting on 10/30/14.
4.4 (preliminary) targets - review next week - MMM 10/27/14
Jeff to distribute this week.
Testing and development with 4.2.22 stable?
Juergen interested in a memo series. Perhaps host on the wiki and publish eternally.
Start anew vs continuing the AIPS++ memo series.
Still testing parallel processing infrastructure w/ MMS.
Problems reported by parallel-go. TBD origin.
Completed JIRA tickets:
CAS-6974 - add buttons to Plot tab in plotms; fixed seg fault reported last week
CAS-6990 - switch between script and GUI clients with 'showgui' argument to plotms
CAS-7014 - selected tools stay selected when iterating through plots
Fixed the bug in MSUVbin mainly this week..a subtle rounding error causing some decorrelation.
Completed ePEP yet again.
Some discussion about weighting in mosaic in general.
Justo González Villalba
fixing a correlation bug - sub-set of correlations would be lost.
Worked on CARTA state.
Able to allocate time last week for first time to CARTA now that Pam is on plotms.
Completed NRAO PEP self evalution after finally receiving guidance from HR on a technical issue.
Spoke with Darrell about package that might be helpful in stats framework. Not sure it would be useful depending on the level of caching needed, although since I have received no input otherwise, the requirement still remains that the caller is responsible for ensuring all datasets are simultaneously available to the stats framework. And anyway, since the stats framework would be most effective if it is located in casacore, and we know how our partners feel about adding dependencies to casacore ... nuff said.
Responded on email thread re: flag shadowing of ALMA antenaas.
Looked at tickets to figure out what I could work on next and finish in a reasonable amount of time. Such tickets are dwindling in number. It would be good if management were to prioritize the top two or three, since many of these are very specialized requests from people not on the CSSC and I am not clear if they should be worked on at all, while others are high cost (long time frame) items.
Discussed interferometry principles with Rob
Fixed polcal_20080224_cband_regression.py which was getting the wrong answers for source polarization (but not actually failing!) because setjy was being run with fluxdensity set but without standard='manual', which is now required.
Responded to some questions from ALMA Polarization SV Casaguide group
Jury duty most of Tuesday (mistrial!)
Working mainly on writing a comprehensive test of weight calibration (w/ and w/out WEIGHT_SPECTRUM) and weight accounting through mstransform/split2
Some diagnostic work on some strange disk (lustre) access behavior (very poor performance) in applycal when running the ngc3256 regression (the dataset of which is serving for my weights test)
Some work on my "EVLA Polarization" talk for the EVLA DR workshop next week
Some work on my "Synthesis Calibration" chapter for the new Summer School text (due Nov ~1)
Began my PEP
completed first version of automated prerelease testing script
worked on generic (unpacked binary+rpm) launcher script
testing group interview
prerelease currently rebuilding, hopefully the last failing test is fixed
Prep for the HPC meeting and follow up on action items. Reading Justo and Sandra's documentation, entering new JIRA tickets, etc.
Maintenance with the helpdesk - updating rpms on casa-dev nodes, casa workstations, repair of machines, updates to launchers for EL6, etc.
Pulling new tests/stables for AOC users. Working on a script to automate the process.
patched casa to build on ubuntu 14.04, still needs some work before upstreaming
built casa in a bare-bone centos 6.5 docker container, updated wiki instruction with issues encountered.
setup personal build farm based on docker + distcc + debug fission, scales pretty good for the slow c++ compiles
Fixed buildCoordinateSystem to correctly handle a default phasecenter
Looked at optical velocity grid issue in spectral grid transformation code a bit more
User support: on setjy and ephemeris data format conversion
Friday NAOJ Meeting
update several SD unit test codes to verify flag hadling