VEGAS Pulsar Project Meeting - Sprint Review: 2017 July 31st 2:00 - 3:00pm ET

Present

• Luke, Ray, Richard, Mark, Ryan,

Sprint Story Overview

• VE-292: Investigate BOF files for possible source of 1-ms spurs (NO UPDATE)
• VE-284: Continue pulsar mode debugging (IN PROGRESS)
• VE-286: Continue debugging 32-channel pulsar mode BOF files (IN PROGRESS)
• VE-257: High time resolution modes (INCOMPLETE)
• VE-290: Reinstate nchan software register in 32 channel modes (NO UPDATE)
• VE-291: Implement fix for 800 MHz BOF file (IN PROGRESS)
• VE-215: Spectral line regression testing (COMPLETE)
• VE-278: Continue regression testing of RC1 (NO UPDATE)
• VE-287: Continue work on pulsar modes for matrix (NO UPDATE)

Story Details

VE-292: Invest BOF files for possible source of 1-ms spurs

• No update

VE-284: Continue pulsar mode debugging

• Coordinator and manager now configure properly and seem to be in sync. Cal-mode scans run properly on vegas-hpc10 using the reg_test installation. However, fold-mode scans are not working because the manager cannot find a tempo installation, which is needed to generate phase predictors. Joe suggested adding an environment variable to gbt.bash but this did not help. The old config file specified relevant paths. Ray will check on this.

VE-286: Continue debugging 32-channel pulsar mode BOF files

• The i1500x0032 mode shows a linear shift in the passband. Ryan investigated this and exonerated config tool and the IF system. Test tone injection confirms that there is a 400 MHz shift in frequency but all LOs are tuned properly. Randy will look more at the BOF file design itself.

VE-257: High time resolution modes

• No update

VE-290: Reinstate nchan software register in 32 channel modes

• No update

VE-291: Implement fix for 800 MHz BOF file

• Luke and Ryan need to double check on the status of this.

VE-215: Spectral line regression testing

• Richard feels it was overall successful. About 25% of the time the VEGAS fits file name is 1-second off from the file names in the coordinator, which breaks the filler. There is also a number of "VEGAS hung in aborting" messages when aborting a scan. While these appear cosmetic it would be good to address them. There are also some remaining issues with the new manager. However, data qualitatively look very good between the "old" VEGAS HPC and the "new" Matrix HPC. Further, all aborts seemed easy to recover from. Further details on wiki. Ray has started working on addressing some of these. Ray think sit should be fairly straightforward to fix the file time/name issues. Note that Joe is out for a few weeks.
• Richard asks what we consider to be "good enough"? Ryan would like to make sure the R3 version is an improvement over the current version when released. The "Manager hung in aborting" and "Arm time in the past" failures seem to be most common. If we can fix one of these we may want to release this version immediately. Richard suggests focusing on this version until end of summer maintenance to utilize as much maintenance time as possible.
• Ray will confirm that there are no issues running guppi_daq and Matrix spectral line spectral line modes.

VE-278: Continue regression testing of RC1

• No update

VE-287: Continue work on pulsar modes for matrix

• No update

VEGAS Pulsar Project Meeting - Sprint Planning: 2017 June 19th 2:00 - 3:00pm ET

Note: The team's "velocity" has been consistent at 11--13 points per sprint. We should aim for 12 points per sprint for planning purposes.

Backlog

• High time resolution modes
• Frequency shift in i1500x0032 mode
• Investigate packet loss in c1500 modes
• Reinstate nchan software register in 32 channel modes

Critical Issues

• Fix remaining pulsar mode bugs in release candidate 1.
• Fix any bugs revealed during spectral line regression tests. Continue regression tests as appropriate.

New Stories

• Continue to test new 24-tap BOF files

-- RyanLynch - 2017-07-31