Researchers previously proposed warp processors, a novel architecture capable of transparently optimizing an executing application by dynamically re-implementing critical kernels within the software as custom hardware circuits in an on-chip FPGA. However, the original warp processor design was primarily performance-driven and did not focus on power consumption, which is becoming an increasingly important design constraint. Focusing on power consumption, we present an alternative low-power warp processor design and methodology that can dynamically and transparently reduce power consumption of an executing application with no degradation in system performance, achieving an average reduction in power consumption of 74%. We further demonstrate the flexibility of this approach to provide dynamic control between high-performance and low-power consumption. Keywords Warp processing, low-power, hardware/software partitioning, dynamically adaptable systems, embedded systems.
Roman L. Lysecky