Reducing dynamic power consumption in FPGAs using precomputation

This paper studies the effectiveness of employing precomputation in reducing dynamic power consumption in commercial off-the-shelf (COTS) FPGAs. Precomputation is a high-level logic optimization technique that lowers power consumption of a design by disabling part of the circuit based on a few relatively simple precomputation conditions. With careful design considerations, the increased logic utilization and its associated power consumption can be justified by the power saving resulted from disabling a much larger part of the design. This fundamental trade-off benefits particularly well from the tile-based structures of modern FPGAs that consist of large number of redundant logic cells. Using the design of a comparator as an example, we study the trade-offs and unique opportunities provided by modern FPGA architectures in employing precomputation as a technique to reduce dynamic power consumption. In our example, 83% of dynamic power from logic, or 43.1% of total dynamic power including routing is reduced with negligible increase in resource consumption.