Data-Intensive Distributed Systems LaboratoryCS Seminar
Date: February 25th, 2026
Time: 12:50pm - 1:45pm
Room: SB111
Dr. Minxuan Zhou
Assistant Professor
Computer Science Department
Illinois Institute of Technology
Talk Title
Accelerating Emerging Applications via Software-Hardware Co-Design
Talk Abstract
Emerging applications are increasingly bottlenecked not by a lack of compute, but by the sheer volume of data they move and the algorithmic complexity they execute. These applications are stretching computing systems in ways that "faster hardware" or "better software" alone cannot fully address. One promising path to achieve gains in performance, efficiency, and usability is software-hardware co-design: jointly shaping algorithms, runtimes, compilers, and architectures around real end-to-end workloads and constraints. In this talk, I will focus on recent works on software-hardware co-design for two fast-moving application domains where co-design has become especially impactful. First, in machine learning, I will present how to accelerate compute-intensive kernels using non-conventional compute architecture with extensive compiler-level kernel mapping optimizations. Second, I will present recent efforts on hardware and software for accelerating cryptography-based privacy-preserving computing.
Speaker Bio
Minxuan Zhou is an Assistant Professor in the Computer Science Department at Illinois Tech since 2024. His research focuses on innovating the hardware architecture and software stack of computing systems for accelerating emerging applications. His research interests include computer architecture, memory systems, compiler, machine learning acceleration, and privacy-preserving computing. Prior to joining Illinois Tech, he was a postdoctoral researcher at the University of California San Diego, where he also got his PhD in 2023.