| Abstract Scope |
To resolve the ever-growing energy consumption challenge in computing in the new information technology (IT) era, correlated materials such as ferromagnets and multiferroics have been employed in microelectronics, providing a promising pathway to develop the next-generation beyond-CMOS microelectronics and bring down the energy dissipation. However, the lack of device-level modeling tools limits the understanding of beyond-Moore devices and systems, and thus adversely affects future advances in microelectronics. I will introduce our efforts in implementing device modeling algorithms that fully leverage modern supercomputers. Our developed numerical tools offer algorithmic flexibility, proven accuracy, portability, and massively GPU-accelerated efficiency that is far beyond the reach of commercial and other research-grade tools. The software currently encompasses electronic, nanomagnetic, ferroelectric, and nanomechanical capabilities, and has been used to design innovative devices, such as compact magnetoelectric spin-orbit logic (MESO) and negative-capacitance field effect transistors (NCFET), and to inspire new devices such as strain-mediated magnetoelectric antennas. |