Expand description
Matrices stored in column-major layout.
Multiplying a column-major matrix by one or more column vectors is fast
due to the way it is implemented in SIMD
(a matrix * vector mutliply is four broadcasts, one SIMD product and three fused-multiply-adds).
Because matrix * matrix and matrix * vector products are fastest with this layout,
it’s the preferred one for most computer graphics libraries and application.