Cryptology ePrint Archive: Report 2021/533

CryptGPU: Fast Privacy-Preserving Machine Learning on the GPU

Sijun Tan and Brian Knott and 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.

Category / Keywords: cryptographic protocols / privacy-preserving machine learning; private inference; private training; MPC; GPU

Original Publication (with minor differences): IEEE Symposium on Security and Privacy 2021

Date: received 22 Apr 2021

Contact author: st8eu at virginia edu,brianknott@fb com,yuant@virginia edu,dwu4@virginia edu

Available format(s): PDF | BibTeX Citation

Version: 20210423:123005 (All versions of this report)

Short URL: ia.cr/2021/533


[ Cryptology ePrint archive ]