This INSPIRE award is partially funded by the Electronics, Photonics, and Magnetics Devices Program in the Division of Electrical, Communications and Cyber Systems in the Directorate for Engineering; the Software and Hardware Foundations program in the Division of Computing and Communication Foundation in the Directorate for Computer & Information Science and Engineering; and the Electronic and Photonic Materials program in the Division of Materials Research in the Directorate for Mathematical and Physical Sciences.
SPINTOP will explore, in detail, a new paradigm of automata computing that is based on a non-Boolean, non-Von Neumann architecture and is enabled by the electrical coupling of Spin Torque Nano-oscillators (STNOs). The project's goal is to build a simple, yet clear demonstration, of how this can achieve significant speedups in pattern recognition to match detailed simulations. This project addresses the 'Big Data' challenge with a non-conventional solution that can compute on the fly. The interdisciplinary nature of the INSPIRE program will make it possible to combine magnetic materials research with device design and test and computer architecture development involving three NSF Programs and faculty from Physics, Materials Science and Electronic and Computer Engineering at University of Virginia (UVa).
Intellectual Merit: SPINTOP will be pushing the frontiers of spintronic devices, circuits and computing architectures. Key to the success of this project will be the demonstration of a novel hybrid STNO that was invented at UVa. Magnetic materials with low magnetization, low damping, high spin polarization, controlled anisotropy are key to the success of the proposed hybrid device. The proposed novel STNO combines the best features of existing devices to provide both pure sine wave output and a large voltage signal. Electrical coupling of arrays of these STNOs will be demonstrated in this project. According to detailed simulations, phase lock will occur when their individual frequencies are close - this phase locking will take place across the array rather than only nearest neighbor locking as has been previously demonstrated using spin wave coupling. This coherent phase locking is the key to the design for an associative memory for pattern recognition. SPINTOP will also develop the modeling tool used to simulate the behavior of the innovative design for STNO arrays. Further development of this tool is key for designing and implementing circuits as it will be integrated with existing codes for performing the modeling, simulation and design of hybrid CMOS-spintronic circuits.
Broader Impacts: SPINTOP, if successful, will have a big impact on the scientific community by creating a new paradigm for computing, especially pattern recognition. This will help solve many of the Big Data applications that are a grand challenge today. However its impact will be much broader; this project will address many of the key issues facing society today. This project is committed to fostering diversity in the workforce and will support at least one student from an underrepresented group for the duration of the project. As the INSPIRE project will fall under the umbrella of the UVA nanoSTAR Institute and the PI is also the director of nanoSTAR, this project will be closely integrated into the outreach programs and benefit from the resources of nanoSTAR. The SPINTOP faculty and students will participate in outreach visits to Virginia high schools to highlight this new technology. SPINTOP faculty and students will also participate in Nanodays, a highly interactive event held each Spring that brings over 650 local students from K-12 and their parents to UVA to experience some of the wonders of nanoscience and technology. These visitors will see a poster highlighting SPINTOP and will be invited to tour the SPINTOP laboratories. SPINTOP along with NanoSTAR will co-sponsor a club for undergraduates called NeXT (for Nano and Emerging Technologies) that hosts frequent seminars with guest speakers and will be a way for students in many of the nano-related disciplines to connect with one another. SPINTOP will leverage programs that support undergraduate research to recruit undergraduates to participate in summer research in the SPINTOP labs. In collaboration with UVA's Center for Diversity in Engineering, opportunities will be provided to several minority senior high school students to experience research in the SPINTOP facilities. SPINTOP will also have an impact on courses at UVA; there will be at least one new class developed under the auspices of this program. It will be a class on automata based computing which is the computing paradigm that the SPINTOP project will use for mapping applications at the system level. Finally, an existing class on Spintronics taught by the PI will be updated to include several lectures on spin torque oscillators and their application to computing.
