Kazi Asifuzzaman is a Research Scientist at Oak Ridge National Laboratory (ORNL). His research focuses on advanced memory systems, emerging computer architectures, and novel computing
paradigms, including neuromorphic computing, analog in-memory computing, and processing-in-memory techniques. He also works at the intersection of quantum computing, high-performance computing, and artificial intelligence, exploring energy-efficient architectures and system-level solutions for next-generation scientific applications.
Presentation Title:
Beyond the Memory Wall: The Evolution of Memory Systems from Storage to Intelligent Computing
Presentation Abstract:
The rapid growth of artificial intelligence, scientific computing, and data-intensive applications is fundamentally reshaping the role of memory systems in modern computing. Traditional memory architectures, built around the Von Neumann model, is increasingly constrained by the memory wall, rising data movement costs, and the unprecedented growth of global data generation. These challenges demand a shift from viewing memory as passive storage toward treating it as an active participant in computation.
This talk presents a unified perspective on the evolution of memory systems, beginning with conventional memory hierarchies and their architectural limitations, progressing through processing-in-memory and analog in-memory computing enabled by emerging technologies such as ReRAM, and highlighting the ReSpike framework for cross-layer evaluation of neuromorphic architectures. The presentation concludes by exploring future storage paradigms, including holographic and DNAbased storage, and discusses how these technologies may converge into a hierarchical intelligent memory ecosystem capable of supporting next-generation AI, high-performance computing, and data-centric scientific discovery.