arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2512.18345cs.CRcs.AR
收藏

Theodosian:深入探讨以内存层次为中心的FHE加速

Theodosian: A Deep Dive into Memory-Hierarchy-Centric FHE Acceleration

Seoul National University(首尔大学)

Wonseok Choi, Hyunah Yu, Jongmin Kim, Hyesung Ji, Jaiyoung Park, Jung Ho Ahn

AI总结:

本文分析了CKKS在现代GPU上的内存层次,发现内核受限于芯片L2缓存,提出Theodosian优化提升缓存效率和运行时性能,实现1.45-1.83倍加速。

摘要:

全同态加密(FHE)使加密数据上的安全计算成为可能,缓解了云计算和边缘环境中的隐私问题。然而,由于其高计算和内存需求,已有大量加速研究在不同硬件平台上进行,尤其是GPU。本文对CKKS这一流行的FHE方案进行了微架构分析,聚焦内存层次,发现主导内核仍受限于芯片L2缓存,尽管其带宽高,暴露了传统离芯片DRAM瓶颈之外的持久内内存墙。进一步发现,整体CKKS吞吐量受限于低的每内核硬件利用率,由内核内并行度不足导致。受这些发现启发,我们引入Theodosian,一组互补的内存感知优化,提升缓存效率并减少运行时开销。Theodosian在代表性CKKS工作负载上比高度优化的基准Cheddar实现了1.45-1.83倍的性能提升。在RTX 5090上,将32768个复数的启动时间从22.1ms减少到15.2ms,并进一步减少到12.8ms,建立目前最好的GPU性能记录。

英文摘要:

Fully homomorphic encryption (FHE) enables secure computation on encrypted data, mitigating privacy concerns in cloud and edge environments. However, due to its high compute and memory demands, extensive acceleration research has been pursued across diverse hardware platforms, especially GPUs. In this paper, we perform a microarchitectural analysis of CKKS, a popular FHE scheme, on modern GPUs. Focusing on the memory hierarchy, we demonstrate that dominant kernels remain bound by the on-chip L2 cache despite its high bandwidth, exposing a persistent inner memory wall beyond the conventional off-chip DRAM bottleneck. Further, we reveal that the overall CKKS throughput is constrained by low per-kernel hardware utilization, caused by insufficient intra-kernel parallelism. Motivated by these findings, we introduce Theodosian, a set of complementary, memory-aware optimizations that improve cache efficiency and reduce runtime overheads. Theodosian achieves 1.45--1.83x performance improvements over a highly optimized baseline, Cheddar, across representative CKKS workloads. On an RTX 5090, we reduce the bootstrapping latency for 32,768 complex numbers from 22.1ms to 15.2ms, and further to 12.8ms with additional algorithmic optimizations, establishing a new state-of-the-art GPU performance to the best of our knowledge.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑