• Starts: 10:00 am on Monday, November 25, 2024
  • Ends: 12:00 pm on Monday, November 25, 2024

ECE PhD Prospectus Defense: Guowei Yang

Title: Domain-Specific Accelerators Using Optically-Addressed Phase Change Memory

Presenter: Guowei Yang

Advisor: Professor Ajay Joshi

Chair: Professor Ayse K. Coskun

Committee: Professor Ajay Joshi, Professor Ayse K. Coskun, Professor Sabrina Neuman

Google Scholar Profile: https://scholar.google.com/citations?user=BdQnapwAAAAJ

Abstract: In recent years, the exponential growth in data generation and the increasing complexity of computational tasks have created a pressing need for more efficient computing solutions. To address this demand, researchers have developed domain-specific accelerators (DSAs) for various applications, including machine learning, combinatorial optimization, and fully homomorphic encryption. However, traditional electric accelerators face significant challenges in performance and energy efficiency. Optically-addressed phase change memory (OPCM) has emerged as a promising solution for accelerating computing because it offers high bandwidth, superior energy efficiency, and process-in-memory capabilities.

In this prospectus, I focus on designing OPCM-based DSAs, applying device-, architecture-, and algorithm-level optimizations. I first present my work on a machine learning accelerator using OPCM. It introduces a system-level design and proposes a novel thresholding and reordering technique to reduce the OPCM programming overhead, achieving up to 65.2× higher throughput than existing photonic accelerators for practical DNN workloads. Then, I present SOPHIE, a Scalable Optical PHase-change memory based Ising Engine. SOPHIE integrates architectural, algorithmic, and device optimizations to address scalability challenges in Ising machines. SOPHIE is 3× faster than the state-of-the-art photonic Ising machines on small graphs and 125× faster than the FPGA-based designs on large problems. SOPHIE alleviates the hardware capacity constraints, offering a scalable and efficient alternative for solving Ising problems. Finally, I propose my future research on fully homomorphic encryption accelerators.

Location:
PHO 339