VEGAS Pulsar Project Meeting - Sprint Review: 2017 August 14th 2:00 - 3:00pm ET

Present

  • Laura, Jason, Ray, Ryan, Natalia

Sprint Story Overview

  • VE-292: Investigate c1500 BOF files for possible source of 1 ms spurs (INCOMPLETE)
  • VE-294: Pulsar mode debugging (IN PROGRESS)
  • VE-257: High time resolution modes (INCOMPLETE)
  • VE-295: Reinstate nchan software register in 32-channel coherent modes (INCOMPLETE)
  • VE-297: Work on issues from regression testing (COMPLETE)
  • VE-281: Continue testing BOF files (INCOMPLETE)
  • VE-293: Continue spectral line regression testing (COMPLETE)
  • VE-296: Continue work on 800 MHz BOF files (COMPLETE)
  • VE-286: Shifts in 32-channel mode

Story Details

VE-292: Investigate c1500 BOF files for possible source of 1 ms spurs

VE-294: Pulsar mode debugging

  • Ryan tested the release candidate version that Richard and Ray were using for regression tests using the full VEGAS system but scans aborted almost immediately. Coherent modes are running into problems in the reg_test environment because they are trying to communicate with banks that are not present, even though they are not selected and the config file says that they should be simulated. Incoherent search mode scans start and run properly but do not exit and the EOF of the fits files is not present. Incoherent modes do seem to be running on reg_test.
  • Ray determined that at least one problem was that each bank was trying to use its own ROACH as its data source, when it should have be the single ROACH used for coherent modes.
    • Hoping to test some of this tomorrow.
  • Ray thinks the issue in search mode may be because the manager is not able to read shared memory, and does not see the DISKSTAT as exiting, hence stopping the scan.

VE-257: High time resolution modes

VE-295: Reinstate nchan software register in 32-channel coherent modes

VE-297: Work on issues from regression testing

  • See notes on regression tests. Most issues were fixed.

VE-281: Continue testing BOF files

  • Was unable to test new builds because reg_test is still not working for pulsar modes and real system was in regression testing.

VE-293: Continue spectral line regression testing

  • See https://safe.nrao.edu/wiki/bin/view/CICADA/CICADAGreenBankSpectrometerCommissioningLogs sessions 109, 110, and 111 for Richard's detailed notes. Some key points reproduced here:
    • Ran for about 12 hours with no VEGAS problems (Astrid seemed to get hung twice, but no sign that this was VEGAS related)
    • "VEGAS hung in aborting" did appear when an abort command was deliberately issued on a few occasions, but quickly clears and appears to be cosmetic.
    • Still occasionally getting messages of the form: “Valon frequency 1500 MHz is not equal to that set”. This was also present in the previous version. This is from an asynchronous process that checks every 30 seconds. It should clear after no more than 30 seconds. This could be fixed by shutting the process down while configuring.
    • Sometimes in an aborted scan, the STRTMJD keyword is missing from the primary header. This doesn’t seem to bother GBTIDL. It may also happen in the old system but this is just speculation. Ray thinks this depends on when the abort is issued.
  • Scientific support staff recommend the following:
    • Look at 2017 projects and determine most commonly used VEGAS modes
    • Perform regression tests of these modes along with setups that are used for RRL (64 spectral windows), KFPA, and ARGUS. As much as possible, use the HI line as a science diagnostic.
    • Ensure there are no issues with fast frequency switching.
  • Barring unforeseen setbacks, we would like to release these modes by the end of summer maintenance.
  • Some CLEO development will also need to be finished.
  • Recommended plan: freeze release candidate no later than Aug 24 for final testing the week of Aug 28.

VE-296: Continue work on 800 MHz BOF files

  • See note for VE-281

VE-286: Shifts in 32-channel mode

  • Randy identified delays in this design that are probably responsible for this. He has implemented a fix and the BOF is building.

VEGAS Pulsar Project Meeting - Sprint Planning: 2017 August 14th 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
  • Investigate packet loss in c1500 modes
  • Reinstate nchan software register in 32 channel modes
  • Continue to test new 24-tap BOF files

Critical Issues

  • Fix remaining pulsar mode bugs in release candidate 1.

New Stories

  • Prepare for final spectral line regression tests by collecting use cases and preparing observing scripts
  • Provide Ron with information needed for CLEO modifications.
  • Test 32-channel mode for frequency shift
 -- RyanLynch - 2017-08-14
Topic revision: r1 - 2017-08-14, RyanLynch
 

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