Towards the least complex time-multiplexed constant multiplication

The multiplication of a variable by a single constant selected from a set of fixed constants at a time, called the time-multiplexed constant multiplication (TMCM), is frequently used in digital signal processing (DSP) systems. Existing algorithms implement the TMCM operation using multiplexers (MUXes), adders/subtractors, and shifts, and reduce its complexity by merging single/multiple constant multiplication graphs and by sharing the basic structures. This paper introduces ARION, that exploits the most common partial terms in the TMCM design on top of the previously proposed DAGfusion algorithm, which merges the single constant multiplication graphs. Experimental results show that ARION obtains significantly better solutions than prominent TMCM methods.