A method and apparatus for optimizing access to memory, wherein the method includes the steps of receiving a first request for access to a memory, receiving at least two additional requests for access to the memory, and determining a first clock overhead associated with the first request for access to the memory. The method further includes the steps of determining an additional clock overhead associated with each of the at least two additional requests for access to the memory in conjunction with the first request, determining a combination of requests that can be processed together using an optimized overhead, and processing the combination of requests as a single request with the optimal overhead.