Embedded Files
This thrust investigates design automation (EDA) challenges due to emerging non-ideal compute technologies.
This thrust investigates design automation (EDA) challenges due to emerging non-ideal compute technologies.
Key Focus Areas
Key Focus Areas
3D Heterogeneous Integration
3D Heterogeneous Integration
How to integrate different CiM devices such as SRAM, ReRAM, and FeFET for efficient DNN training?
Thermal-aware Neural Layer Mapping
Thermal-aware Neural Layer Mapping
How to map CNN neural layers in 3D manycore system such that the impact of temperature on DNN accuracy due to NVM non-idealities is minimized?
Heterogeneous Operation Units (OUs)
Heterogeneous Operation Units (OUs)
How to mitigate challenges such as IR-drop and conductance drift in NVM-based CiM crossbars without compromising energy efficiency and performance?
Page updated
Google Sites
Report abuse