A Streaming Data Processing Architecture Based on Lookup Tables

Processing in memory (PIM) is a new computing paradigm that stores the function values of some input modes lookup table (LUT) and retrieves their when similar are encountered (instead performing online calculations), which an effective way to save energy. In era Internet Things, processing massive data generated by front-end requires low-power real-time processing. This paper investigates energy-efficient architecture based on phase-change (PCM). replaces logical-based calculations with LUT lookups minimize power consumption operation latency. order improve efficiency lookup, RISC-V instruction set has included extended stream transmission instructions. Finally, system validated hardware simulation, performance fast Fourier transform (FFT) application evaluated. The proposed effectively improves execution reduces flow operations.