Add Additional Buffer Variants to Delay and Tapped Delay with Block Size > 1

[cyarp]

The current plan is to keep the existing logic for block size == 1. The delay was implemented to support several different buffering schemes and element ordering (earliest first, latest first).

For the initial implementation of sub-blocking, only a specific subset of these options will be implemented for block size > 1. The other options can be added later if further investigation of buffering schemes is needed.