2020 11 23 Open NEST Developer Video Conference

previous | list | next

Agenda

1. Welcome
2. Review of NEST User Mailing List
3. Project team round
4. In-depth discussion
   • Compiler variants vs. MPI implementations
   • OSS plans on Travis CI for NEST and NESTML

Mailing list

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

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
   • working on numerical issues of the Siegert neurons
   • Travis discussion (see in-depth)
2. PyNEST
   • closing some inactive issues with insufficient information and no replies anymore.
3. Kernel
   • ModelsModule refactoring from SLI function to plain nest::register_builtin_models(), being discussed.
   • Timers being finalized for last barriers being removed, CI needs to rerun.
4. Installation
   • no major updates
   • discussed Fedora libpython linking #1630
5. Infrastructure
   • Travis is up and working again
   • fixed the warning reporting and included ignore-list for unsolvable SLI warnings. now at zero warnings, build will fail if you introduce new ones. Look into the Travis build log, the summary contains the list of files that contains the warnings.
   • need review of the CI pipeline documentation #1837
6. Documentation
   • closed ~10 documentation related PRs in the last weeks
   • NEST Server documentation updated, expect PR soon.
   • build system refactored to unify build and doc system. Requires more testing/review. #1843
7. EBRAINS
   • nothing to report
8. (Feature) Random number generation
   • started to update documentation, RNG guide, etc.
   • need to run some larger benchmarks
9. (Feature) Automated Testing
   • nothing to report
10. (Feature) Extension Module System
    • see kernel
    • ConnGen refactoring done and reviewed #1830, works since MacOS fix is in.
    • SLI dependencies are now all gone
    • planning a hack-a-day for implementing first ideas of new extension module system.

In-depth discussion

Compiler variants vs. MPI implementations

• investigating performance comparison to 2.14
• @akurth saw SetStatus() calls not terminating with large enough number of neurons (368 for serial, 210 for parallel)
• works with gcc, but mpicc does not, intel compilers neither.

Comments:

• maybe bisect between 2.14 and master, may not be a compiler problem
• try plain intel compiler without MPI
• if point of fix is found, back-port could be possible.

Proposed topic: OSS plans on Travis CI for NEST and NESTML

• Travis was bought, now changing payment plans
• now credits for build jobs (translating to minutes of build time)
• open-source projects can change plan for different resource allocation
  • applied for NESTML and for NEST
  • not clear what the OSS plan would actually be
  • https://blog.travis-ci.com/2019-01-23-travis-ci-joins-idera-inc points out that:

    Project must not be sponsored by a commercial company or organization (monetary or with employees paid to work on the project)

• alternatives are coming up everywhere, investigating current possibilities.
  • GitHub actions
  • GitLab runners
  • Jenkins
  • …
• sunset of travis-ci.org till end of the year.

=> investigate options