Parallelism, Performance, and Energy-Efficiency Tradeoffs for In Situ Sensor Data Processing

Abstract

The in situ processing of vast amounts of data, available intermittently in networks of sensors, motivates investigation of means for achieving high performance when required, but ultralow-power dissipation when idle. One approach is the use of embedded multiprocessor systems, leading to tradeoffs between parallelism, performance, energy-efficiency, and cost. To evaluate these tradeoffs, and to gain insight for future system designs, this letter presents the design, implementation, and evaluation of a miniature, energy-scalable, 24-processor module, L24, for compute-intensive in situ sensor data processing tasks. The platform provides idle power dissipation over an order of magnitude lower than systems employing a monolithic processor of equivalent performance, while dynamic power dissipation remains competitive. Taking into account both application computation and interprocessor communication demands, it is shown that there may exist an optimum operating voltage that minimizes either time-to-solution, energy usage, or the energy-delay product. This optimum operating point is formulated analytically, calibrated with system measurements and instruction-level microarchitectural simulation, and evaluated for the hardware platform and application presented.