Repository navigation
ENH: GPU Tutorial 4 meshing (~27x); split ProcessContours - #151
Merged
Merged
Conversation
|
Important Review skippedToo many files! This PR contains 138 files, which is 38 over the limit of 100. To get a review, reduce the PR to 100 files or fewer by splitting it into smaller PRs or changing its base branch. Upgrade to a paid plan to raise the limit. This review couldn't start because sufficient usage credits or metered capacity aren't available. Add credits or update usage-based reviews in the billing tab, then retry. ⚙️ Run configuration
⛔ Files ignored due to path filters: 89
📒 Files selected for processing: 138
You can disable this status message by setting the 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. Comment |
Speed up Tutorial 4 (Duke Heart labelmap to VTK, "full" mode) about 27x on the full cohort (20,569 s to 754 s), with geometry matching the previous output to within 0.1 mm RMSE. Breaking changes (see docs/developer/migration_next.md; conversion via utils/migrate_process_contours.py): - ProcessContours is split into ProcessSurfaceMesh, ProcessVolumetricMesh and the new public ProcessMeshKernels (GPU closest point, inside test, tet quality, distance maps, blur, label-vote resample). - extract_contours and extract_watertight_surface are removed; extract_label_surfaces is the single watertight extractor, and neighboring labels share their walls. - WorkflowConvertImageToVTK accepts a labelmap, reduces surfaces with ACVD, can build tetrahedra, and no longer returns segmentation_masks. - ProcessUSD stage methods take an open Usd.Stage; open_stage and save_stage are new. - EvaluateMovementLung scores NVSegmentCTMRI lobe ids. Packaging and environment: - cuda12/cuda13 extras use cupy[ctk], so CuPy loads pip-installed CUDA libraries instead of a CUDA_PATH toolkit whose cuBLAS conflicted with torch's (CUBLAS_STATUS_NOT_INITIALIZED). - torchaudio is dropped from the extras, docs, CI and the worktree setup script: it is unused, and no build matches current torch. - README and quickstart lead with the prebuilt torch-scatter install. Tooling: - Ruff config consolidated into pyproject.toml (.ruff.toml removed; the never-applied strict rule list is gone). The hook is bumped to v0.16.6 and its new default rules are fixed repo-wide: deliberate catch-alls are marked noqa with a reason, silent passes use contextlib.suppress, and the rest are narrowed. - The mypy hook runs in the project venv, so it type-checks against the installed torch, warp and pyvista; the type errors this exposed are fixed, and stale mypy overrides are pruned. Tests: - The contour-mesh fixture builds its labelmap through test_labelmaps instead of skipping when the cached file is missing. - Stale "pytest harness" notes are removed from the GPU training tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Tutorial 1, 3, 4, 5 and 6 heart baselines were generated by --create-baselines while CuPy could not load its CUDA headers and torchaudio mismatched torch, so they do not match current output. Regenerate them once the environment is stable. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
aylward
force-pushed
the
perf/t04-full-runtime
branch
from
October 20, 2026 00:32
f89722e to
e6d24b6
Compare
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
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.
Speed up Tutorial 4 (Duke Heart labelmap to VTK, "full" mode) about 27x on the full cohort (20,569 s to 754 s), with geometry matching the previous output to within 0.1 mm RMSE.
Breaking changes (see docs/developer/migration_next.md; conversion via utils/migrate_process_contours.py):
Packaging and environment:
Tooling:
Tests: