/// Naive non-cooperative matmul without tiling that can be very fast on small matrices.
pub mod naive;
pub mod vecmat_plane_parallel;
pub mod vecmat_unit_perpendicular;
pub mod double_buffering;
pub mod double_unit;
pub mod interleaved;
pub mod ordered_double_buffering;
pub mod simple;
pub mod simple_unit;
pub mod specialized;
pub mod vecmat_innerproduct;
mod base;
mod selector;
#[cfg(feature = "tensor-matmul")]
pub(crate) mod gemv_ordered;
pub use base::*;
pub use selector::*;