# Objective - People have reported bounding volumes being slower than their existing solution because it doesn't use SIMD aligned types. ## Solution - Use `Vec3A` internally for bounding volumes, accepting `Into<Vec3A>` wherever possible - Change some code to make it more likely SIMD operations are used. --- ## Changelog - Use `Vec3A` for 3D bounding volumes and raycasts ## Migration Guide - 3D bounding volumes now use `Vec3A` types internally, return values from methods on them now return `Vec3A` instead of `Vec3` |
||
|---|---|---|
| .. | ||
| src | ||
| Cargo.toml | ||