One of the bottlenecks is residue computers has been in translating residue numbers into polynomial. I have made an improvement in the speed of that process by inventing a method for translating small numbers into polynomial faster than large ones. Prior to my work, the speed of the translation process was dependant on the number of residue digits the machine could process, but was independent of the data being processed.
Thus a machine with a wider data bus would take a longer time to translate a small number than would a machine with a smaller data bus. I discovered that it is possible to achieve the capability of translating small numbers faster than larger ones simply by adding an and gate to the residue processor. This means that small numbers can be divided faster than larger ones. In a polynomial processor it takes a barrel shifter to achieve this, and a barrel shifter is exponentially more complicated than the simple and gate circuit needed to achieve the same results in a residue system.