Second Homomorphic Encryption Computing Performance Optimization Forum – Hardware Accelerators

Forum Overview The Second Homomorphic Encryption Computing Performance Optimization Forum – Homomorphic Encryption Hardware Accelerators takes place on October 26, 13:30‑17:30 in the Summer Garden – Rural Club Conference Room 1. The event continues the discussion from last year, focusing on core challenges, multi‑disciplinary research, and innovative solutions for accelerating homomorphic encryption.

Highlights • Ongoing in‑depth discussions on performance optimization and knowledge updates. • Participation of high‑level experts, including contributors from the previous CCF‑Ant Financial Research Fund. • Emphasis on cross‑domain research, encouraging innovative thinking and interdisciplinary collaboration. • Presentation of constructive ideas and suggestions to address performance bottlenecks and promote practical deployment.

Schedule The forum features five sessions:

1. General Homomorphic Encryption Accelerator – Speaker: Zhang Mingzhe (Ant Technology Research Institute)

2. Homomorphic Encryption Algorithm Compilation – Speaker: Bian Song (Beihang University)

3. Performance Trade‑offs and Architectural Extensions in Fully Homomorphic Encryption Hardware Design – Speaker: Wang Zhongfeng (Sun Yat‑sen University)

4. FHE Soft‑Hardware Acceleration Design Space Exploration – Speaker: Wang Xiaohang (Zhejiang University)

5. Near‑Data Computing in Privacy‑Preserving Computation – Speaker: Zhang Jie (Peking University)

Speaker Introductions and Abstracts

Zhang Mingzhe – Deputy Director, Computing Systems Lab, Ant Technology Research Institute. Report Title: General Homomorphic Encryption Accelerator. Abstract: Describes a unified accelerator architecture that supports multiple homomorphic encryption schemes and their conversion, achieving performance beyond existing single‑scheme solutions while addressing design challenges.

Bian Song – Associate Professor, School of Cyberspace Security, Beihang University. Report Title: Homomorphic Encryption Algorithm Compilation. Abstract: Explores automatic compilation of general computations into encrypted operators, focusing on multi‑level intermediate representations, Boolean‑circuit generation tools, and optimization techniques for encrypted computation.

Wang Zhongfeng – Dean and Professor, School of Integrated Circuits, Sun Yat‑sen University. Report Title: Performance Trade‑offs and Architectural Extensions in Fully Homomorphic Encryption Hardware Design. Abstract: Discusses joint software‑hardware optimization for programmable, reconfigurable homomorphic encryption, presenting design trade‑offs and theoretical insights for co‑design.

Wang Xiaohang – Professor, School of Cybersecurity, Zhejiang University. Report Title: FHE Soft‑Hardware Acceleration Design Space Exploration. Abstract: Proposes a graph‑based model to explore chip design space for FHE applications, formulating an optimization problem to minimize execution time under cost and power constraints.

Zhang Jie – Assistant Professor, School of Computer Science, Peking University. Report Title: Near‑Data Computing in Privacy‑Preserving Computation. Abstract: Highlights the limitations of von Neumann architectures for federated learning and confidential computing, and presents a near‑data computing solution that tightly couples software‑hardware co‑design to overcome physical data movement bottlenecks.

Additional Information The forum is part of CNCC2024 (China Computer Congress), which runs from October 24‑26 in Hengdian, Zhejiang. CNCC focuses on “Developing New‑Quality Productivity, Computing Leading the Future” and gathers Turing Award winners, academicians, top scholars, and industry leaders.