Skip to content

Conversation

@david-christiansen
Copy link
Contributor

This PR adds a Radar setup so we can track the performance of doc-gen4 builds.

The benchmarks are:

  • own-docs measures the time and maxrss for building the docs for doc-gen itself. Here, it is a convenient stand-in for a moderately-sized Lean project with a moderate number of depenencies.

  • mathlib-docs measures the time and maxrss for building Mathlib's docs. We can use it to ensure that from-scratch Mathlib documentation builds don't get smaller.

This PR adds a Radar setup so we can track the performance of doc-gen4
builds.

The benchmarks are:

* `own-docs` measures the time and maxrss for building the docs for
  doc-gen itself. Here, it is a convenient stand-in for a
  moderately-sized Lean project with a moderate number of depenencies.

* `mathlib-docs` measures the time and maxrss for building Mathlib's
  docs. We can use it to ensure that from-scratch Mathlib
  documentation builds don't get smaller.
@Garmelon
Copy link

!bench

@leanprover-radar
Copy link

leanprover-radar commented Jan 16, 2026

Benchmark results for 7e278bc against 5b5f81c are in! @Garmelon

Runs (1🟥)
  • 🟥 main exited with code 2

@Garmelon
Copy link

!bench

@leanprover-radar
Copy link

Benchmark results for 7e278bc against 5b5f81c are in! @Garmelon

Runs (1🟥)
  • 🟥 main exited with code 1

@Garmelon
Copy link

Hm, it fails to get a git remote at some point in the build, and later other things fail due to heartbeat limits.

@david-christiansen
Copy link
Contributor Author

!bench

@leanprover-radar
Copy link

leanprover-radar commented Jan 16, 2026

Benchmark results for 7361fd8 against 5b5f81c are in! @david-christiansen

Runs (1🟥)
  • 🟥 main exited with code 255

@david-christiansen
Copy link
Contributor Author

!bench

@leanprover-radar
Copy link

leanprover-radar commented Jan 16, 2026

Benchmark results for 24375ae against 5b5f81c are in! @david-christiansen

Runs (1🟥)
  • 🟥 main exited with code 1

@Garmelon
Copy link

!bench

@leanprover-radar
Copy link

Benchmark results for 24375ae against 5b5f81c are in! @Garmelon

No significant changes detected.

@david-christiansen
Copy link
Contributor Author

!bench

@leanprover-radar
Copy link

leanprover-radar commented Jan 17, 2026

Benchmark results for c366b8c against 5b5f81c are in! @david-christiansen

No significant changes detected.

This better isolates the actual work of generating the docs site, and
measures the correct target for Mathlib.
@david-christiansen
Copy link
Contributor Author

!bench

@leanprover-radar
Copy link

leanprover-radar commented Jan 17, 2026

Benchmark results for d560bd5 against 5b5f81c are in! @david-christiansen

No significant changes detected.

@Garmelon
Copy link

Seems to work, I'd merge.

@david-christiansen david-christiansen marked this pull request as ready for review January 17, 2026 21:02
@david-christiansen david-christiansen merged commit 837f89a into leanprover:main Jan 17, 2026
1 check passed
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.

3 participants