Paper 2022/1222

Homomorphic Encryption on GPU

Ali Şah Özcan, Sabanci University
Can Ayduman, Sabanci University
Enes Recep Türkoğlu, Sabanci University
Erkay Savaş, Sabanci University
Abstract

Homomorphic encryption (HE) is a cryptosystem that allows secure processing of encrypted data. One of the most popular HE schemes is the Brakerski-Fan-Vercauteren (BFV), which supports somewhat (SWHE) and fully homomorphic encryption (FHE). Since overly involved arithmetic operations of HE schemes are amenable to concurrent computation, GPU devices can be instrumental in facilitating the practical use of HE in real world applications thanks to their superior parallel processing capacity. This paper presents an optimized and highly parallelized GPU library to accelerate the BFV scheme. This library includes state-of-the-art implementations of Number Theoretic Transform (NTT) and inverse NTT that minimize the GPU kernel function calls. It makes an efficient use of the GPU memory hierarchy and computes 128 NTT operations for ring dimension of $2^{14}$ only in $176.1~\mu s$ on RTX~3060Ti GPU. To the best of our knowlede, this is the fastest implementation in the literature. The library also improves the performance of the homomorphic operations of the BFV scheme. Although the library can be independently used, it is also fully integrated with the Microsoft SEAL library, which is a well-known HE library that also implements the BFV scheme. For one ciphertext multiplication, for the ring dimension $2^{14}$ and the modulus bit size of $438$, our GPU implementation offers $\mathbf{63.4}$ times speedup over the SEAL library running on a high-end CPU. The library compares favorably with other state-of-the-art GPU implementations of NTT and the BFV operations. Finally, we implement a privacy-preserving application that classifies encrpyted genome data for tumor types and achieve speedups of $42.98$ and $5.7$ over a CPU implementations using single and 16 threads, respectively. Our results indicate that GPU implementations can facilitate the deployment of homomorphic cryptographic libraries in real world privacy preserving applications.

Metadata
Available format(s)
PDF
Category
Implementation
Publication info
Preprint.
Keywords
Lattice Based Cryptography Homomorphic Encryption Number Theoretic Transform NTT GPU Secure Computation
Contact author(s)
alisah @ sabanciuniv edu
canayduman @ sabanciuniv edu
eturkoglu @ sabanciuniv edu
erkays @ sabaciuniv edu
History
2022-11-17: last of 3 revisions
2022-09-15: received
See all versions
Short URL
https://ia.cr/2022/1222
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2022/1222,
      author = {Ali Şah Özcan and Can Ayduman and Enes Recep Türkoğlu and Erkay Savaş},
      title = {Homomorphic Encryption on GPU},
      howpublished = {Cryptology ePrint Archive, Paper 2022/1222},
      year = {2022},
      note = {\url{https://eprint.iacr.org/2022/1222}},
      url = {https://eprint.iacr.org/2022/1222}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.