Technique used to speed up processing time; by splitting banks of memory each with x microseconds access time and accessing each one in sequence 1/n-th of a cycle later, a reference to memory can be had every x /n microseconds; this speed is achieved at the cost of hardware complexity.