DRAMatic Speedup: Accelerating HE Operations on a Processing-in-Memory System

TL;DR

DRAMatic accelerates homomorphic encryption operations on UPMEM PIM hardware, achieving over 300x speed-up through arithmetic optimizations, but faces data transfer and multiplication performance challenges.

Contribution

It introduces DRAMatic, a novel PIM-based system with optimized HE operations, significantly improving speed and energy efficiency over prior implementations.

Findings

DRAMatic achieves 334x speed-up over previous HE implementations on UPMEM PIM.

DRAMatic reduces the gap between UPMEM PIM and Microsoft SEAL in runtime and energy efficiency.

Data transfer overhead and limited multiplication performance are current bottlenecks.

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. We present DRAMatic, which implements operations foundational to HE on UPMEM PIM -- a programmable general-purpose PIM system developed by UPMEM. 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. It achieves a 334 times speed-up compared to previous HE implementations on UPMEM PIM. We also 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 Microsoft SEAL and HE implementations on UPMEM PIM. However, we also show that DRAMatic is currently constrained by data transfer overhead and limited multiplication performance on UPMEM PIM hardware. Finally, we discuss potential hardware extensions to UPMEM PIM.