E-ReCON: An Energy- and Resource-Efficient Precision-Configurable Sparse nvCIM Macro for Conventional and Spiking Neural Edge Inference
E-ReCON introduces E-ReCON, a 16 Kb energy-efficient digital compute-in-memory (DCIM) macro using a compact 3T1R ReRAM bitcell for edge AI. It eliminates ADC/DAC overhead and supports both conventional and spiking neural networks with high accuracy, achieving up to 419 TOPS/W efficiency.