DRAMatic Speedup: Accelerating HE Operations on a Processing-in-Memory System
Application Category
📝 Abstract
Homomorphic encryption (HE) is a promising technology for confidential cloud computing, as it allows computations on encrypted data. However, HE is computationally expensive and often memory-bound on conventional computer architectures. Processing-in-Memory (PIM) is an alternative hardware architecture that integrates processing units and memory on the same chip or memory module. PIM enables higher memory bandwidth than conventional architectures and could thus be suitable for accelerating HE. In this work, we present DRAMatic, which implements operations foundational to HE on UPMEM's programmable, general-purpose PIM system, and evaluate its performance. DRAMatic incorporates many arithmetic optimizations, including residue number system and number-theoretic transform techniques, and can support the large parameters required for secure homomorphic evaluations. To compare performance, we evaluate DRAMatic against Microsoft SEAL, a popular open-source HE library, regarding both runtime and energy efficiency. The results show that DRAMatic significantly closes the gap between UPMEM PIM and Microsoft SEAL. However, we also show that DRAMatic is currently constrained by UPMEM PIM's multiplication performance and data transfer overhead. Finally, we discuss potential hardware extensions to UPMEM PIM.
Problem
Research questions and friction points this paper is trying to address.
Homomorphic Encryption
Processing-in-Memory
Memory-bound
Computational Overhead
Hardware Acceleration
Innovation
Methods, ideas, or system contributions that make the work stand out.
Processing-in-Memory
Homomorphic Encryption
Residue Number System
Number-Theoretic Transform
UPMEM
🔎 Similar Papers
No similar papers found.
