A RRAM-based FPGA for Energy-efficient Edge Computing

As the data processing shift from centralized cloud to edge computing it demands hardware systems with data processing capability at ultra-low power. FPGA being one of the most popular solution for edge computing systems because of the flexibility.

An RRAM based FPGA architecture is proposed for edge computing with low power consumption and high performance. RRAM is a non-volatile memory technology having low power consumption and high-speed operation.

There are no major architectural differences with respect to general FPGA only SRAM based primitive blocks are replaced by RRAM based circuits for non-volatility. For analysis open-source tool OpenFPGA is used with modification for supporting RRAM based FPGA architecture.

The proposed RRAM based model when compared to SRAM based counterpart shows 8% improvement in area, 22% improvement in delay and 16% improvement in power.