Traditional computing faces a critical bottleneck hindering its evolution toward more efficient and powerful systems. Dr. Hassan Najafi, an assistant professor in UL 鶹ýapp’s School of Computing and Informatics and a leading expert in computational science, developed an innovative solution poised to reshape the landscape of computing as we know it.
Najafi recently received one of the National Science Foundation’s most prestigious awards – the Faculty Early Career Development Award. This award acknowledges a faculty member’s potential as a researcher and a teacher.
As part of the award, Najafi will receive a nearly $600,000 grant over several years to fund his research on conventional computing systems and improvements to these systems.
Conventional computing systems struggle with the costly and time-consuming transfer of data between memory and processing units. This inefficiency not only drains energy but also hampers processing speed, particularly in data-driven applications. Termed the "Von Neumann bottleneck," this limitation stifles the potential development of groundbreaking computational systems.
Enter In-Memory Computation (IMC), also known as Processing in Memory (PIM). This paradigm shift offers a promising solution to the data movement dilemma. However, the reliability issues inherent in emerging memory devices pose significant challenges, often leading to computation errors.
To address these challenges, Najafi’s research proposes combining the strengths of Unary Computing (UC) – an emerging paradigm offering simplicity in executing complex operations and high immunity to noise – with IMC to form UCoM, a groundbreaking framework for implementing UC in memory.
“UCoM boasts unparalleled performance, energy efficiency, and robustness across various tasks. Its highly parallel nature scales seamlessly with computation size, supporting a wide array of essential arithmetic operations entirely in memory,” said Najafi. “UCoM promises substantially higher performance and significantly lower energy consumption compared to conventional computing systems.”
Moreover, this project extends its impact beyond technical innovation. The project includes a comprehensive educational and outreach plan, engaging students from diverse backgrounds in emerging computing technologies. It aims to train the next generation of computational scientists while disseminating project outcomes for free public access. Furthermore, partnerships with industry stakeholders seek to commercialize these innovations, driving real-world impact and adoption.
“Dr. Najafi’s research including UCoM represents a significant step towards the realization of high-performance computing systems with brain-like efficiency and power. By embedding unary computing in memory, this groundbreaking framework heralds a new era of energy efficient and fast computational systems,” said Dr. Azmy Ackleh, dean of the Ray P. Authement College of Sciences.
Najafi is the 10th recipient of the prestigious CAREER award in the College of Sciences. Ackleh noted this demonstrates the college’s continued success in recruiting high-caliber faculty and its contributions to making the University a Research 1 top-tier university. Only 3% of public and private institutions in the United States have earned this designation by the Carnegie Classification of Institutions of Higher Education.
The NSF is an independent federal agency. It provides funds for nearly a quarter of all federally supported research American colleges and universities conduct. NSF first presented the CAREER Award in 1995.
Photo caption: Dr. Hassan Najafi, assistant professor in UL 鶹ýapp’s School of Computing and Informatics and recipient of the National Science Foundation’s prestigious Faculty Early Career Development Award. Courtesy photo