Skip to content

perf(ci): nightly perf aggregator + Numba social/PSM accumulators#1

Open
razinkele wants to merge 10 commits into
mainfrom
perf/nightly-aggregator-numba
Open

perf(ci): nightly perf aggregator + Numba social/PSM accumulators#1
razinkele wants to merge 10 commits into
mainfrom
perf/nightly-aggregator-numba

Conversation

@razinkele

Copy link
Copy Markdown
Owner

Summary: Fix births/deaths accounting; add Numba accumulators, perf CI and nightly aggregator. See docs/PERFORMANCE.md for details.

…f: add Numba accumulators and social aggregator; tests, docs
@coderabbitai

coderabbitai Bot commented Jan 25, 2026

Copy link
Copy Markdown

Caution

Review failed

An error occurred during the review process. Please try again later.


Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands and usage tips.

@github-actions

Copy link
Copy Markdown

Perf check

  • ticks: 500
  • porpoise: 500
  • elapsed: 1.70s
  • baseline: 10
  • threshold: 14.5
  • regression: NO

@github-actions

Copy link
Copy Markdown

Perf check

  • ticks: 500
  • porpoise: 500
  • elapsed: 1.67s
  • baseline: 10
  • threshold: 14.5
  • regression: NO

…; move energy accumulation out of _update_psm
@github-actions

Copy link
Copy Markdown

Perf check

  • ticks: 500
  • porpoise: 500
  • elapsed: 3.78s
  • baseline: 10
  • threshold: 14.5
  • regression: NO

@github-actions

Copy link
Copy Markdown

Perf check

  • ticks: 500
  • porpoise: 500
  • elapsed: 1.74s
  • baseline: 10
  • threshold: 14.5
  • regression: NO

razinkele pushed a commit that referenced this pull request Jun 12, 2026
Verified all findings against code:
- #3 CRITICAL: JSON loader defaults impact=170 -> every Kattegat ship forced to
  170dB override -> JOMOPANS would be inert. Loader also ignores real type/length
  (mostly OTHER, hardcoded 100m). Added loader fixes (impact default, type->VesselClass
  map, read length) + vessel-class validation.
- #1: only ONE site (population.py:2718); main.py:2102 is deter_max_distance, not a
  threshold -> corrected.
- #2: JAX gate threads through tick_jax.py jax_tick_energy (not just jax_kernels.py:843);
  jax_tick_energy takes BOTH combined (BMR/reporting) + turbine-only (dispersal) arrays.
  Cython path needs no change; scalar porpoise.deter() noted as deferred.
- Added 3 existing-test updates (test_jax_tick:1248, test_dispersal:337, test_integration:163).
DEPONS parity of approach confirmed (band 12, speed-0, turbine-only).
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant