VEGAS Pulsar Project Meeting - Sprint Review: 2017 July 17th 2:00 - 3:00pm ET

Present

  • Laura, Jason, Randy, Joe

Sprint Story Overview

  • VE-276: Connect 10Gb Ethernet switch for test environment (COMPLETE)
  • VE-279: Pulsar mode debugging (IN PROGRESS)
  • VE-278: Continue Regression Test Preparation of RC1 (COMPLETE)
  • VE-277: Test 32-channel coherent dedispersion BOF files (IN PROGRESS)
  • VE-280: Investigate c0800x2048 bandpass issues
  • VE-283: Start work on pulsar modes for matrix (R3)
  • VE-281: Continue testing BOF files (COMPLETE)
  • VE-282: Work with Justin on Manager

Story Details

VE-276: Connect 10Gb Ethernet switch for test environment

  • Hooked up and appears to be working.

VE-279: Pulsar mode debugging

  • Config tool validates pulsar modes without issue and seems to set parameters correctly in the Coordinator, but parameters are not always propagating down to the Manager under certain circumstances. Specifically, it seems like the Coordinator and Manager do not sync up until a parameter has been manually set in the Manager. Even a conform params and prepare immediately following a configure does not force a change in the Manager state. The Manager seems to remain in a default state. Joe has looked into this in Ray's absence. The Manager seems to have been written in such a way that until parameters have been set twice, it uses default values from the vegas.conf file. Ryan notes that this behavior is not seen in Justin's M&C branch, i.e. parameters are set the first time correctly. Joe thinks the problem is in the dependency calculation for obs_mode in the Manager.

VE-278: Continue Regression Test Preparation of RC1

  • Spectral line modes seem ready for full regression tests.

VE-277: Test 32-channel coherent dedispersion BOF files

  • Justin's branch of the Manager would not configure for 32-channel coherent modes. Incoherent modes seemed to mostly check out, except for what appears to be a frequency shift in the i1500x0032 mode. We suspect this is an issue with IF configuration, but Ryan did not have a chance to further investigate this. Randy reviewed the BOF file design and found no obvious issues that would cause the frequency shift.
  • Randy notes that some software registers have changed/been eliminated. This definitely could cause problems. Randy will build a version with dummy registers to present the expected registers to the Manager.

VE-280: Investigate c0800x2048 bandpass issues

  • Luke identified a problem in the previous BOF file and has a new build that should be ready for testing soon.

VE-283: Start work on pulsar modes for matrix (R3)

  • Joe has started work on this. Currently fixing error handling.

VE-281: Continue testing BOF files

  • All new BOF files have been tested initially. Still outstanding issues with c0800x2048, and packet loss in i1500x2048, i1500x4096, and c1500 modes.

VE-282: Work with Justin on Manager

  • Incoherent 32-channel modes configure but not coherent 32-channel modes.

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

  • Complete testing 32-channel BOF files
  • Test any new BOF files

Critical Issues

  • Complete spectral line regression tests
  • Fix remaining pulsar mode bugs in release candidate 1

New Stories

  • Try to improve dropped packet rate in c1500 modes
  • Quantify improvements in new 24-tap PFB BOF files.
  • Build 32-channel coherent BOF files with new dummy registers.
-- RyanLynch - 2017-07-17
Topic revision: r2 - 2017-07-17, 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