ci: get the test + benchmark workflows green (VMEC2000-from-source, fork-PR benchmark guard)#583
Closed
krystophny wants to merge 6 commits into
Closed
ci: get the test + benchmark workflows green (VMEC2000-from-source, fork-PR benchmark guard)#583krystophny wants to merge 6 commits into
krystophny wants to merge 6 commits into
Conversation
krystophny
changed the title
krystophny
force-pushed
the
ci-bazel-build-fixes
branch
from
31fbbde to
07e9fdb
Compare
krystophny
changed the title
The 'Compare benchmark result' step uses github-action-benchmark with comment-on-alert and the GITHUB_TOKEN, which is read-only for pull requests from forks -> 'Resource not accessible by integration'. Gate that step on the PR coming from the same repo so fork PRs still run the benchmarks but skip the write-back instead of failing.
krystophny
force-pushed
the
ci-bazel-build-fixes
branch
from
07e9fdb to
c9cb71f
Compare
krystophny
changed the title
Contributor
Author
|
@jurasic-pf could you review this one first, independently of the differentiable-VMEC++ series? It is CI-only and standalone (no source/build changes), and it gets the test + benchmark workflows green again on
Merge order: this PR first. Once it lands on |
The pinned vmec-0.0.6 cp310 wheel was f90wrapped against numpy 1.x. Under the numpy 2.x that the test env now resolves, importing it dies in the f90wrap array interface (f90wrap_vmec_input__array__rbc: 0-th dimension must be fixed to 2 but got 4), so test_ensure_vmec2000_input_from_vmecpp_input could never actually run on CI (and is currently red on main too, where the wheel's runtime libs are not even installed). Build VMEC2000 from upstream source with current f90wrap, which produces numpy-2-compatible bindings. The recipe mirrors SIMSOPT's own CI (hiddenSymmetries/VMEC2000, cmake/machines/ubuntu.json). An explicit 'import vmec' check in the install step surfaces any remaining problem here rather than as a confusing test failure.
With VMEC2000 built from current upstream source, the compatibility test runs for the first time and hits vmecpp indata fields that have no counterpart in the legacy VMEC2000 INDATA namelist (e.g. free_boundary_method), which raised AttributeError. The test explicitly checks only the common subset, so guard the lookup with hasattr and skip fields VMEC2000 does not have, instead of enumerating them one by one.
krystophny
force-pushed
the
ci-bazel-build-fixes
branch
from
1604813 to
3c42253
Compare
Build the VMEC2000 wheel from the exact commit the vmec-0.0.6 wheel was built from (hiddenSymmetries/VMEC2000 @ 728af8bd6c79, the 'works with numpy 2.0' merge), and cache it keyed on that SHA. Same pinned, known-good source as the old wheel, but f90wrapped against the current numpy so it imports under numpy 2; only the first CI run pays the build cost, the rest restore the cached wheel. Addresses the request to keep a pinned, fast VMEC2000 install.
jurasic-pf
reviewed
Jun 14, 2026
jurasic-pf
reviewed
Jun 15, 2026
| # cache it, so only the first run pays the build cost. Pinned and | ||
| # known-good like the old wheel, just numpy-2-correct. | ||
| sudo apt-get install -y libopenmpi-dev libnetcdff-dev libscalapack-mpi-dev \ | ||
| libhdf5-dev libhdf5-serial-dev libfftw3-dev libopenblas-dev ninja-build |
Collaborator
There was a problem hiding this comment.
Unclear what FFTW is needed for, I'd prefer to avoid GPL licensed code where possible
Contributor
Author
There was a problem hiding this comment.
elminated. was just part of the "all you need for vmec2000 build" but actually not needed.
jurasic-pf
approved these changes
Jun 15, 2026
jurasic-pf
left a comment
jurasic-pf
left a comment
Collaborator
There was a problem hiding this comment.
Thanks! This really went unnoticed for way too long!
Co-authored-by: Philipp Jurašić <166746189+jurasic-pf@users.noreply.github.com>
VMEC2000 at the pinned commit 728af8b references no FFTW anywhere; its build needs only MPI, SCALAPACK, NETCDF (Fortran+C) and BLAS/LAPACK. VMEC++ uses vendored FFTX codelets, not system fftw3. HDF5 arrives transitively via libnetcdff-dev, so the explicit libhdf5 packages are redundant too.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash, not the tag.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash for Clang >= 21.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash for Clang >= 21.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash for Clang >= 21.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash for Clang >= 21.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash for Clang >= 21.
krystophny
added a commit
to itpplasma/vmecpp
that referenced
this pull request
Jun 15, 2026
…mit pin Bring this stack branch up to the corrected CI baseline (from proximafusion#583/proximafusion#564): - tests.yaml: build VMEC2000 from the pinned source commit and cache the wheel; drop the unused FFTW/HDF5 dev packages. - benchmarks.yaml: skip the result upload on fork PRs (read-only token). - test_simsopt_compat.py: skip vmecpp-only INDATA fields. - CMakeLists: pin abseil to the 20260107.1 commit hash for Clang >= 21.
Contributor
Author
|
Closing in favor of #564, which now carries all of these CI fixes plus
The review threads here are addressed in #564. Consolidating into one PR |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Scope: CI only, standalone. No source or build-system changes. Independent of
the differentiable-VMEC++ PR stack (#564-#582); please review and merge this one
first so those PRs inherit green CI from
main.What it fixes
1. VMEC2000 compatibility test (
test_ensure_vmec2000_input_from_vmecpp_input).The pinned
vmec-0.0.6cp310 wheel was f90wrapped against numpy 1.x. The testenv now resolves numpy 2.x, under which importing that wheel dies in the f90wrap
array interface:
So the test could never actually run. (It is red on
maintoo: the wheel'sruntime libs - libnetcdff, scalapack, openblas - are not installed there, and
pytest 9's
importorskiponly skipsModuleNotFoundError, so a missing.sois a hard failure rather than a skip.)
Fix: build VMEC2000 from upstream source with current
f90wrap, which producesnumpy-2-compatible bindings. The recipe mirrors SIMSOPT's own CI
(
hiddenSymmetries/VMEC2000,cmake/machines/ubuntu.json). No fork needed. Anexplicit
import vmeccheck in the install step surfaces any remaining problemthere rather than as a confusing test failure.
2. Benchmark workflow on fork PRs. The result-upload step needs a writable
GITHUB_TOKEN; on fork PRs the token is read-only, so it failed with"Resource not accessible by integration". Gate that step to same-repo PRs; the
benchmark itself still runs.
Verification
Failing before (ubuntu-22.04, py3.10):
import vmecchain broke onlibnetcdff -> libscalapack -> libopenblas, then on the f90wrap/numpy-2
rbcValueError above.Passing after: VMEC2000 builds from source,
import vmec->VMEC2000 import OK,and the compatibility test runs. Benchmark workflow no longer errors on the
upload step for fork PRs.