Lune

HPCA2025Top-tier venue

WarpDrive: GPU-Based Fully Homomorphic Encryption Acceleration Leveraging Tensor and CUDA Cores

Guang Fan, Mingzhe Zhang, Fangyu Zheng, Shengyu Fan, Tian Zhou, Xianglong Deng, Wenxu Tang, Liang Kong, Yixuan Song, Shoumeng Yan

Abstract

The application of Fully Homomorphic Encryption (FHE) is rapidly gaining traction as a means to maintain data confidentiality while performing computations on encrypted data. Given the accessibility and computational power, GPUs hold promise for significantly accelerating FHE operations. However, existing GPU-based acceleration solutions face several formidable challenges, notably the extensive occurrence of pipeline stalls induced by memory access and suboptimal harnessing of GPU hardware. This paper presents WarpDrive, a comprehensive framework for GPU-based FHE acceleration. Through sophisticated computation decomposition and fine-grained memory access design, WarpDrive significantly reduces the number of instructions by 73%\mathbf{7 3 \%} and pipeline stalls by 86%\mathbf{8 6 \%} compared to the state-of-the-art solution. Additionally, WarpDrive features a framework that supports the concurrent utilization of CUDA Cores and Tensor Cores within the NTT operation, for the first time, achieving performance that surpasses that of any single type of processing unit. Furthermore, we fully exploit the intra-ciphertext parallelism to elevate both computation and memory utilization, achieving up to 2.12×2.12 \times improvements without the need for ciphertext batching. Experimental results demonstrate that our optimizations highly enhance the performance of homomorphic operations. On an NVIDIA A100 GPU, WarpDrive achieves a throughput of 1218 KOPS for NTT and 305 KOPS for homomorphic multiplication, outperforming the state-of-the-art GPU solution (TensorFHE) by factors of 13.4×13.4 \times and 3.5×3.5 \times, respectively. For the specific FHE workload, even under a much smaller batch size, our approach achieves 2.8×2.8 \times the performance of TensorFHE.

Ask Lune

Know where this paper stands in minutes, not an afternoon.

Lune can read the top-tier papers that cite it and the ones closest to it, and tells you what it claims, what came after and where it falls short, each point cited to the paper it comes from. No skimming, no AI slop.

5 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

its abstract · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Fenc2: Unifying Data Packing for Efficient Private Inference via Convolution and Architecture-Aware Fragment Encoding cites Cheddar: A Swift Fully Homomorphic Encryption Library Designed for GPU Architectures
  • Leveraging ASIC AI Chips for Homomorphic Encryption cites Cheddar: A Swift Fully Homomorphic Encryption Library Designed for GPU Architectures
  • Cerium: A Multi-GPU Framework for Terabyte-Scale Encrypted Inference cites Leveraging ASIC AI Chips for Homomorphic Encryption
  • Cerium: A Multi-GPU Framework for Terabyte-Scale Encrypted Inference cites Cheddar: A Swift Fully Homomorphic Encryption Library Designed for GPU Architectures
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers5

  1. Cheddar: A Swift Fully Homomorphic Encryption Library Designed for GPU ArchitecturesWonseok Choi, Jongmin Kim, Jung Ho AhnASPLOS 2026 · 6 citations
  2. Leveraging ASIC AI Chips for Homomorphic EncryptionJianming Tong, Tianhao Huang, Jingtian Dang, Leo de Castro et al.HPCA 2026 · 2 citations
  3. Fenc2: Unifying Data Packing for Efficient Private Inference via Convolution and Architecture-Aware Fragment EncodingRan Ran, Zhaoting Gong, Nuo Xu, Yuanchao Xu et al.ISCA 2026
  4. BitDecoding: Unlocking Tensor Cores for Long-Context LLMs with Low-Bit KV CacheDayou Du, Shijie Cao, Jianyi Cheng, Luo Mai et al.HPCA 2026
  5. Cerium: A Multi-GPU Framework for Terabyte-Scale Encrypted InferenceSiddharth Jayashankar, Joshua Kim, Michael B. Sullivan, Wenting Zheng et al.SOSP 2026

Related papers

  1. HyperDrive: Hierarchical Exploitation of Memory Efficiency for GPU-Based FHE AccelerationGuang Fan, Yi Chen, Lei Chen, Liang Kong et al.ISCA 2026
  2. TensorFHE: Achieving Practical Computation on Encrypted Data Using GPGPUShengyu Fan, Zhiwei Wang, Weizhi Xu, Rui Hou et al.HPCA 2023 · 90 citations
  3. MNEMOS: A GPU-Based TFHE Acceleration Framework with Memory Access OptimizationJunyi Zhang, Xianglong Deng, Yi Chen, Guang Fan et al.ISCA 2026
  4. Neo: Towards Efficient Fully Homomorphic Encryption Acceleration using Tensor CoreDian Jiao, Xianglong Deng, Zhiwei Wang, Shengyu Fan et al.ISCA 2025 · 18 citations
  5. Strix: An End-to-End Streaming Architecture with Two-Level Ciphertext Batching for Fully Homomorphic Encryption with Programmable BootstrappingAdiwena Putra, Prasetiyo, Yi Chen, John Kim et al.MICRO 2023 · 27 citations
Dusk over the sea between two cliffs drawn in fine vertical lines