Resolution matrix-vector multiple speedups

[sbailey]

See desihub/fastspecfit#181 which "Reimplement[s] enough of desispec.resolution.Resolution to perform matrix-vector operations, which is all we do with it, with a significant speedup via Numba compared to the scipy.sparse implementation."

Consider whether any of that should be back-ported to desispec, while still retaining the full functionality of the current Resolution object.