2021 05 10 Open NEST Developer Video Conference

previous | list | next

Agenda

1. Welcome
2. Review of NEST User Mailing List
   • created #2033, discussed different variants of "interactive simulations"
3. Project team round
4. In-depth discussion
   • Interactively debugging GitHub Actions failures
   • Discussion of STDP synapse unit testing #1840
5. Join the NEST Conference

Mailing list

• https://nest-simulator.org/mailinglist

Project team round

Here we discuss topics that need broader attention, for example questions that came up but are outside a single project's scope, larger planned changes/PRs that affect all teams or pending work that is blocked by external factors.

1. Models / NESTML
   • NESTML 4.0 now ready for release
   • (see in depth)
2. PyNEST
   • nothing to report
3. Kernel
   • #1338 is merged!
4. Installation
   • issue #1959, needs MacOS testing, looking for helping hands for pyenv tests
5. Infrastructure
   • nothing to report
6. Documentation
   • looking for reviewers on various PRs
   • currently implementing results from discussion on style guide
7. EBRAINS
   • Looking for high-level support contact persons
8. (Feature) Automated Testing
   • see in-depth
9. (Feature) Extension Module System
   • work in progress

In-depth discussion

Interactively debugging GitHub Actions failures

• issue last week with failing unit tests, but not reproducible locally
• GHA instanciates VM(s), which can be instructed to stay online after tests completed.
  • adding a stages that includes https://github.com/mxschmitt/action-tmate allows to SSH into the VM.
  • uses tmate.io service
  • potentially exposes tokens in the environment of the VM
  • remove the extra stages before creating a pull-request!

Discussion of STDP synapse unit testing #1840

• STDP synapse unit-testing, checking the weight updates with own/separate reference code

• see also #1840 revealed that plasticity mechanisms did not take precise timing into account correctly.

• Question: apply this fix to all other synapse models?

  • could break various current unit tests, where numeric results are compared to specific reference values.
  • e.g. test/test_stdp_multiplicity.py : L.223
  • weight_events where also not taking into account precise spike times (values correctly, but different time?)
    • weights were calculated and reported on grid (consistent), but STDP was not tested with precise neurons.

• Question: PR done? or need to do all other synapses too? -- Keep together in this PR to find the context in future.

• Comment: need to simplify and unify internal access to event times, not e.get_stamp().get_ms() - e.get_offset()

  • revisit at next hackathon!

Join the NEST Conference 2021

• Conference registration and abstract submission are open!