what it's contract says needs to happen (and what existing code expects).

non-one based strides.

interfaces.  Also changed the default behavior when the stride isn't 1.  Now
the filters will be applied only to the "valid" part of the image.

to expose it in the final layer interface.

known at set() time.

values were used on some pixel types.

allow you to find out which device owns the memory inside these objects.

streambufs before exiting.

when multiple GPUs are used.

than raw X and Y matrices.