eprint.iacr.org will be offline for approximately an hour for routine maintenance at 11pm UTC on Tuesday, April 16. We lost some data between April 12 and April 14, and some authors have been notified that they need to resubmit their papers.

Paper 2021/533

CryptGPU: Fast Privacy-Preserving Machine Learning on the GPU

Sijun Tan, Brian Knott, Yuan Tian, and David J. Wu

Abstract

We introduce CryptGPU, a system for privacy-preserving machine learning that implements all operations on the GPU (graphics processing unit). Just as GPUs played a pivotal role in the success of modern deep learning, they are also essential for realizing scalable privacy-preserving deep learning. In this work, we start by introducing a new interface to losslessly embed cryptographic operations over secret-shared values (in a discrete domain) into floating-point operations that can be processed by highly-optimized CUDA kernels for linear algebra. We then identify a sequence of "GPU-friendly" cryptographic protocols to enable privacy-preserving evaluation of both linear and non-linear operations on the GPU. Our microbenchmarks indicate that our private GPU-based convolution protocol is over 150x faster than the analogous CPU-based protocol; for non-linear operations like the ReLU activation function, our GPU-based protocol is around 10x faster than its CPU analog. With CryptGPU, we support private inference and private training on convolutional neural networks with over 60 million parameters as well as handle large datasets like ImageNet. Compared to the previous state-of-the-art, when considering large models and datasets, our protocols achieve a 2x to 8x improvement in private inference and a 6x to 36x improvement for private training. Our work not only showcases the viability of performing secure multiparty computation (MPC) entirely on the GPU to enable fast privacy-preserving machine learning, but also highlights the importance of designing new MPC primitives that can take full advantage of the GPU's computing capabilities.

Metadata
Available format(s)
PDF
Category
Cryptographic protocols
Publication info
Published elsewhere. Minor revision. IEEE Symposium on Security and Privacy 2021
Keywords
privacy-preserving machine learningprivate inferenceprivate trainingMPCGPU
Contact author(s)
st8eu @ virginia edu
brianknott @ fb com
yuant @ virginia edu
dwu4 @ virginia edu
History
2021-04-23: received
Short URL
https://ia.cr/2021/533
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2021/533,
      author = {Sijun Tan and Brian Knott and Yuan Tian and David J.  Wu},
      title = {CryptGPU: Fast Privacy-Preserving Machine Learning on the GPU},
      howpublished = {Cryptology ePrint Archive, Paper 2021/533},
      year = {2021},
      note = {\url{https://eprint.iacr.org/2021/533}},
      url = {https://eprint.iacr.org/2021/533}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.