Low Overhead Online Checkpoint for Intermittently Powered Non-volatile FPGAs

Abstract

Energy harvesting is an attractive way to power future IoT devices since it can eliminate the need for battery or power cables. However, harvested energy is intrinsically unstable. While FPGAs have been widely adopted in various embedded systems, it is hard to survive unstable power since all the memory components in FPGA are based on volatile SRAMs. The emerging non-volatile memory based FPGAs provide promising potentials to keep configuration data during power outages. However, few works have considered implementing efficient runtime intermediate data checkpoint on non-volatile FPGAs. To realize accumulative computation under intermittent power on FPGA, this paper proposes a low-cost design, FC-FPGAs, which utilizes "scan-chain like" flip-flops to track intermediate data. Instead of keeping all on-chip intermediate data, FC-FPGA only targets on necessary data that is labeled by off-line analysis and identified by an on-line tracking circuit. The evaluation shows that compared with state-of-the-art, FC-FPGA can realize accumulative computing and significantly reduce computation time and energy over a wide range of unstable power traces.