GeoT Optimization 4/4: single-pass context+local extraction (halve radius_search)#1744
GeoT Optimization 4/4: single-pass context+local extraction (halve radius_search)#1744coreyjadams wants to merge 6 commits into
Conversation
Add a sync-free aliasing check (_same_coords) so the context projector detects when spatial_coords and geometry are views of the same storage (common after collate unsqueeze) and runs each per-scale processor once via extract_context_and_local instead of separately in extract_context_features and extract_local_features, halving radius_search calls. Falls back to the two-pass path when inputs differ. Enable include_local_features in the surface recipe config and add coverage.
|
Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
peterdsharpe
left a comment
There was a problem hiding this comment.
Looks good and makes sense! Nice optimization here.
Greptile SummaryThis PR adds a single-pass
Important Files Changed
Reviews (1): Last reviewed commit: "Cleaning up the fast path radius search" | Re-trigger Greptile |
|
/ok to test b0b9b80 |
PhysicsNeMo Pull Request
Add a sync-free aliasing check (_same_coords) so the context projector detects when spatial_coords and geometry are views of the same storage (common after collate unsqueeze) and runs each per-scale processor once via extract_context_and_local instead of separately in extract_context_features and extract_local_features, halving radius_search calls. Falls back to the two-pass path when inputs differ. Enable include_local_features in the surface recipe config and add coverage.
Description
Checklist
Dependencies
Review Process
All PRs are reviewed by the PhysicsNeMo team before merging.
Depending on which files are changed, GitHub may automatically assign a maintainer for review.
We are also testing AI-based code review tools (e.g., Greptile), which may add automated comments with a confidence score.
This score reflects the AI’s assessment of merge readiness and is not a qualitative judgment of your work, nor is
it an indication that the PR will be accepted / rejected.
AI-generated feedback should be reviewed critically for usefulness.
You are not required to respond to every AI comment, but they are intended to help both authors and reviewers.
Please react to Greptile comments with 👍 or 👎 to provide feedback on their accuracy.