Feature request: `Octavian.batched_mul` and `Octavian.batched_mul!`

[DilumAluthge]

NNlib.jl has the following nice functions:

Function	Description
batched_mul(A, B) -> C	Batched matrix multiplication. Result has C[:,:,k] == A[:,:,k] * B[:,:,k] for all k. If size(B,3) == 1 then instead C[:,:,k] == A[:,:,k] * B[:,:,1], and similarly for A.
batched_mul!(C, A, B) -> C and batched_mul!(C, A, B, α=1, β=0)	In-place batched matrix multiplication, equivalent to mul!(C[:,:,k], A[:,:,k], B[:,:,k], α, β) for all k. If size(B,3) == 1 then every batch uses B[:,:,1] instead.

Of course, under the hood, all of these functions call LinearAlgebra.mul!. It would be nice if Octavian could have its own batched_mul and batched_mul! that calls our own pure Julia matmul routines.

Octavian.batched_mul and Octavian.batched_mul! would be public functions, i.e. we would document them and put them in the manual. But we probably won't export them, just so we don't clash with NNlib.

[chriselrod]

It's worth exploring if some sort of repacking could improve performance, in particular for large batches of small matrices.