Uncriticality-Directed Low-Power Instruction Scheduling

Intelligent mobile information devices require low-power and high-performance processors. In order to reduce energy consumption with maintaining computing performance, we proposed to utilize information regarding instruction criticality. Every functional unit in our processor has different latency and energy consumption. Only critical instructions are executed in power-hungry units, and the total energy consumption can be reduced. While we have studied several techniques to identify critical instructions for years, we have not found any technique that achieves both simplicity and high accuracy to identify. In this paper, we propose to exploit uncriticality rather than criticality. Only uncritical instructions are executed in power-efficient units, and energy consumption can be reduced. Our simulation results show approximately 10% energy reduction.