Power optimisation of combinational circuits by input transformations

An approach is presented for power minimisation of CMOS combinational circuits based on transforming power-expensive inputs to functionally equivalent low-power inputs. It is assumed that the environment information is available in the form of either word-level statistics or a long representative input sequence. From the environment information, power-efficient input transformations are derived that will save power for any input stream that satisfies the environment properties. A transform module is synthesised that is prepended to the primary inputs of the module. The proposed approach can be applied at the system level by propagating environment information to the primary outputs of the system and then deriving transform circuits for each module. As an example of real-world application, results are presented for a DCT module in a JPEG compression chip with different input environments (e.g. face recognition dataset and weather forecasting). Experimental results are highly encouraging.