ALGORITHMS FOR LOW-POWER AND HIGH-SPEED FIR FILTER REALIZATION USING DIFFERENTIAL COEFFICIENTS

Most realizations of FIR filters use the coefficients directly to comp ute the convolution with the input data. We present a set of new algor ithms for low power and high-speed realization of FIR filters. The alg orithms use various orders of differences between coefficients for com puting the convolution, Also the results of computations are stored an d reused, thus requiring more storage and storage accesses, These tech niques result in a reduction in the net computations necessary per con volution as compared to directly using the coefficients. It is shown a nalytically that this computational reduction at the cost of more stor age can result in a lower net computational energy dissipated and an e nhanced speed, These algorithms are applied to an example FIR filter t o quantify the energy savings and speedup.