[BUG FIX] Revert fix safe GJK fallback causing dramatic performance drop.#2716
Conversation
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
duburcqa
changed the title
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 954a223015
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| count = 1 | ||
| dot_max = dot | ||
| # FIXME: This fix destroys performance (-25% total runtime FPS on benchmarks). Disabling for now... | ||
| # dot_max = dot |
There was a problem hiding this comment.
In _func_count_supports_mesh, removing dot_max = dot means dot_max stays at -1e20, so the if dot > dot_max branch is taken on every loop iteration and count is repeatedly reset to 1. As a result, mesh support counting can never return multiple valid supports, which disables the multi-support perturbation path used by GJK (see gjk.py around if num_supports > 1) and can reintroduce degenerate-simplex failures for mesh contacts.
Useful? React with 👍 / 👎.
1 participant
No description provided.