Research lab: iCAS Lab @SJTU, advised by Prof. Xinfei Guo
Research interest: compute-in-memory, processing-in-memory, reconfigurable computing
Cross-Layer Reconfigurable SRAM-CIM
Compute-in-memory (CIM) is an effective approach to deal with “Memory Wall“ issue in CPU development nowadays. Targeting at edge AI scenarios, we aim at finding an energy efficient and low area consumption CIM solution for the accelerator designs. We also want to integrate more functionalities/operators on a CIM macro, which can be configured flexibly according to needs, to have better performance than general-purpose accelerators and lower design cost than fully-customized accelerators.