Forging the Future of AI and Security through Memory-centric Computing

Pursuing his Ph.D. in Electrical and Computer Engineering, Purab Sutradhar is at the forefront of developing cutting-edge computing solutions for data-centric and AI applications through his research on memory-centric computing.

Memory-centric computing is a revolutionary concept that makes parallel computing faster, and more cost-effective and energy-efficient by performing computations within or near the memory of a computing device. Under the guidance of Dr. Amlan Ganguly, Purab’s research is aimed at developing novel in/near-memory processing and dataflow techniques within DRAM with an edge in performance and energy efficiency of computing. His design solutions span numerous application domains, including low-power AI acceleration, cryptography, and cyber-physical systems. His groundbreaking research activities have led to peer-reviewed publications in prestigious journals and conferences, as well as a US patent.

Purab believes that memory-centric computing will be at the forefront of digital computing in the near future and will spur a paradigm shift in data-center architectures and intelligent systems. Currently, he is focused on developing hyper-scale, memory-centric architectural solutions aimed at optimized processing of generative AI and cybersecurity applications.

Connecting his research with his passion for teaching, Purab has developed a new course on Memory-centric Architectures for RIT’s Computer Engineering program. This graduate elective, set to be taught in the Spring 2024 semester, aims to translate Purab’s research insights into a fertile learning environment to nurture future researchers.

To learn more about Purab’s research, see the following selected publications:

1. P.R. Sutradhar, S. Bavikadi, M. A. Indovina, S. Dinakarrao, A. Ganguly, “FlutPIM: A Look-up Table-based Processing in Memory Architecture with Floating-Point Computation Support for Deep Learning Applications,” in Proceedings of the Great Lakes Symposium on VLSI (GLSVLSI), 2023  https://doi.org/10.1145/3583781.3590313
2. P.R. Sutradhar, S. Bavikadi, S. Dinakarrao, M. A. Indovina, A. Ganguly, “3DL-PIM: A Look-up Table oriented Programmable Processing in Memory Architecture based on the 3-D Stacked Memory for Data-Intensive Applications,” in IEEE Transactions on Emerging Topics in Computing (TETC), 2023 https://doi.org/10.1109/TETC.2023.3293140
3. P.R. Sutradhar, K. Basu, S. Dinakarrao, A. Ganguly, “An ultra-efficient look-up table based programmable processing in-memory architecture for data encryption” in 39th Int. Conference on Computer Design (ICCD), 2021 https://doi.org/10.1109/ICCD53106.2021.00049