VEGAS Pulsar Project Meeting - Sprint Review: 2017 March 20th 2:00 - 3:00pm ET

Present

• Richard, Mark, Jason, Joe, Ryan, Natalia, Laura, Ray

Sprint Story Overview

• VE-187: Troubleshoot bad Stokes U and V data (IN PROGRESS)
• VE-191: Continue troubleshooting manager stability issues (IN PROGRESS)
• VE-114: Complete fiber testing for regression environment (DONE)
• VE-195: Rebaseline VEGAS plan for quarterly planning (DONE)
• VE-198: Continue troubleshooting issues on simulator (IN PROGRESS)
• VE-196: Finalize decisions on SRBS ownership (DONE)
• VE-185: Regression test of rebased Manager code (INCOMPLETE)
• VE-194: Continue work on spectral line PFB component (IN PROGRESS)
• VE-159: Troubleshoot data visualization tools (INCOMPLETE)
• VE-197: Continue work on high time resolution modes (INCOMPLETE)
• VE-202: Create scripts for merging multi-bank data (DONE)

Story Details

VE-187: Troubleshoot bad Stokes U and V data

• Randy found a potential issue with FPGA delay blocks that are used to obtain timing closure during BOF file builds. For designs that use a lot of resources these could cause significant differences in signal paths between X and Y polarizations, which could cause the observed delays.
• Randy is building BOF files using a different scheme to obtain timing closure that will result in more predictable behavior. Will test this week.

VE-191: Continue troubleshooting manager stability issues

• Ray is testing manager behavior when switching between spectral line modes. He can change, initialize and prepare.
• Data taken in modes 1-3 but there are problems starting scans in modes that use GPU's for DSP.

VE-198: Continue trouble shooting issues on simulator

• In parallel with VE-191

VE-114: Complete fiber testing for regression environment

• Done

VE-195: Rebaseline VEGAS plan for quarterly planning meeting

• Done. See planning notes below.

VE-196: Finalize decision on SRBS ownership

• SRBS is indeed owned by Tim Bastian. Tim says that he may want the system back at some point but for now we can use it. If he asks for it back, we will have to discontinue use.
• We decided to make the recently built "test environment" the "regression environment" as originally outlined in Richard's project plan. SRBS will become a development environment for exclusive use by the software division. The ideal system will need a second GPU-enabled HPC. Laura will talk to computing about existing HPCs that we may be able to use.

VE-185: Regression test of rebased Manager code

• Waiting for completion of VE-191

VE-194: Continue work on spectral line PFB component

• Working on book keeping for blanking and accumulation buffers
• Revised data structures
• Updated packet generator for blanking and switching signals
• Performed some work on accumulators

VE-159: Troubleshoot data visualization tools

• No updates

VE-197: Continue work on high time resolution modes

• No updates

VE-202: Create scripts for merging multi-bank data

• New story
• Ryan had some time to work on scripts for automatically combining coherent and incoherent data taken in multi-bank mode.
• Works with search and fold-mode data for continuous frequency coverage

Outstanding issues

M&C System

• Ray is making sure that spectral line modes behave as expected with rebased code. This is necessary before any regression tests and release.

Stokes U/V Issues

• Randy has some new BOF file designs to test. Dave MacMahon has also been providing some suggestions. We will try looking at ADC snap block data to see if the problem exists there.

VEGAS Pulsar Project Meeting - Sprint Planning: 2017 February 20th 2:00 - 3:00pm ET

Rebaselining Discussion

• Ray thinks M&C could be ready for spectral line regression testing in the next week.
• With luck, pulsar modes will work as they used to, but this is TBD
• New target for release of pulsar modes to observers is May
• Aim for release of LBW pulsar modes in August

Backlog

• Geting M&C working on simulator
• Data visualization
• M&C regression testing
• Stokes U/V issues
• Continue work on matrix
• High time resolution modes

Critical Issues

• M&C system
• Stokes U/V issues

New Stories

• VE-191: VPM Manager stability issues
• VE-198: Get M&C working on simulator
• VE-185: Regression tests in spectral line mode
• VE-187: Investigate Stokes U/V issues
• VE-159: Troubleshoot data visualization tools
• VE-197: Finish checks of data indexing
• VE-203: Clone /home/sim for regression test environment