This is much easier for a user to deal with. It really shouldn't make any difference for performance, and in fact, may enhance it (the old implementation was using copyto!). In situations where it would make a difference, like for insanely large pulses, there's always the option of using a non-allocating functor.
This is much easier for a user to deal with. It really shouldn't make any difference for performance, and in fact, may enhance it (the old implementation was using
copyto!
). In situations where it would make a difference, like for insanely large pulses, there's always the option of using a non-allocating functor.