[Resource Topic] 2022/1321: cuZK: Accelerating Zero-Knowledge Proof with A Faster Parallel Multi-Scalar Multiplication Algorithm on GPUs

Welcome to the resource topic for 2022/1321

Title:
cuZK: Accelerating Zero-Knowledge Proof with A Faster Parallel Multi-Scalar Multiplication Algorithm on GPUs

Authors: Tao Lu, Chengkun Wei, Ruijing Yu, Yi Chen, Li Wang, Chaochao Chen, Zeke Wang, Wenzhi Chen

Abstract:

Zero-knowledge proof (ZKP) is a critical cryptographic protocol, and it has been deployed in various privacy-preserving applications such as cryptocurrencies and verifiable machine learning. Unfortunately, ZKP has a high overhead on its proof generation step, which consists of several time-consuming operations, including large-scale matrix-vector multiplication (MUL), number-theoretic transform (NTT), and multi-scalar multiplication (MSM) on elliptic curves. Currently, several GPU-accelerated implementations of ZKP have been developed to improve its performance. However, these existing GPU designs do not fully unleash the potential of GPUs. Therefore, this paper presents cuZK, an efficient GPU implementation of ZKP with the following three optimizations to achieve higher performance. First, we propose a new parallel MSM algorithm and deploy it in cuZK. This MSM algorithm is well adapted to the high parallelism provided by GPUs, and it achieves nearly perfect linear speedup over the Pippenger algorithm, a well-known serial MSM algorithm. Second, we parallelize the MUL operation, which is lightly disregarded by other existing GPU designs. Indeed, along with our self-designed MSM parallel scheme and well-studied NTT parallel scheme, cuZK achieves the parallelization of all operations in the proof generation step. Third, cuZK reduces the latency overhead caused by CPU-GPU data transfer (DT) by 1) reducing redundant data transfer and 2) overlapping data transfer and device computation with the multi-streaming technique. We design a series of evaluation schemes for cuZK. The evaluation results show that our MSM module provides over 2.08× (up to 2.63×) speedup versus the state-of-the-art GPU implementation. cuZK achieves over 2.65× (up to 4.47×) speedup on standard benchmarks and 2.18× speedup on a GPU-accelerated cryptocurrency application, Filecoin.

ePrint: https://eprint.iacr.org/2022/1321

See all topics related to this paper.

Feel free to post resources that are related to this paper below.

Example resources include: implementations, explanation materials, talks, slides, links to previous discussions on other websites.

For more information, see the rules for Resource Topics .