1 paper
Chenhao Xue, Yukun Wang, An Guo +11
SRAM-based compute-in-memory (CIM) offers high computational density and energy efficiency for deep neural network (DNN) accelerators, but its limited capacity causes on/off-chip d…