17. Approximate Systems: Synergistically Approximating Sensing, Computing, Memory, and Communication Subsystems for Energy Efficiency

Emerging application domains exhibit the property of intrinsic error resilience that enables new avenues for energy optimization of computing systems, namely the introduction of a small amount of approximations during system operation in exchange for substantial energy savings. Almost all prior work in the area of approximate computing has focused on individual subsystems of a computing system, e.g., the computational subsystem or the memory subsystem. Since they focus only on individual subsystems, these techniques are unable to exploit the large energy-saving opportunities that stem from adopting a full-system perspective and approximating multiple subsystems of a computing platform simultaneously in a coordinated manner. Towards this end, this chapter introduces the concept of an Approximate System that performs joint approximations across different subsystems, leading to significant energy benefits compared to approximating individual subsystems in isolation. We use the example of a smart camera system that executes various computer vision and image processing applications to illustrate how the sensing, memory, processing, and communication subsystems can all be approximated synergistically. The approximate smart camera system was implemented using an Altera Stratix IV GX FPGA development board, a Terasic TRDB-D5M 5 Megapixel camera module, a Terasic RFS WiFi module, and a 1 GB DDR3 DRAM SODIMM module. Experimental results obtained using six application benchmarks demonstrate that the proposed full-system approximation methodology achieves significant energy savings of 1.8 × to 5.5 × on average over individual subsystem-level approximations for minimal (<1%) application-level quality loss.