pub fn compose_rotation_scheduled(left: &Mat3I16, right: &Mat3I16) -> Mat3I16Expand description
Compose two Q12 rotation/scale matrices with the GTE’s scheduled MVMVA path.
The left matrix is loaded into the rotation registers and each column of
right is transformed in turn. Translation is forced to zero, so the
result is exactly the rotation/scale product. This clobbers the live GTE
rotation and translation state; callers normally load the returned scene
matrix immediately afterward.