Module simd 

SIMD-friendly batch math operations.

When the simd feature is enabled, these functions structure computations to encourage auto-vectorization by the compiler. On wasm32 with -C target-feature=+simd128, LLVM will use SIMD instructions.

Functions

batch_cross

Batch cross products: compute cross(a[i], b[i]) for parallel arrays.

batch_distance_sq

Batch squared distances between corresponding point pairs.

batch_dot

Batch dot products: compute dot(a[i], b[i]) for parallel arrays.

batch_normalize

Batch normalize: normalize an array of vectors.

batch_transform_points

Batch transform: apply a Mat4 to an array of points.