IACR News item: 24 September 2023
Shiyu Shen, Hao Yang, Wangchen Dai, Lu Zhou, Zhe Liu, Yunlei Zhao
ePrint Report
Homomorphic Encryption (HE) enhances data security by facilitating computations on encrypted data, opening new paths for privacy-focused computations. The Brakerski-Fan-Vercauteren (BFV) scheme, a promising HE scheme, raises considerable performance challenges. Graphics Processing Units (GPUs), with considerable parallel processing abilities, have emerged as an effective solution.
In this work, we present an in-depth study focusing on accelerating and comparing BFV variants on GPUs, including Bajard-Eynard-Hasan-Zucca (BEHZ), Halevi-Polyakov-Shoup (HPS), and other recent variants. We introduce a universal framework accommodating all variants, propose optimized BEHZ implementation, and first support HPS variants with large parameter sets on GPUs. Moreover, we devise several optimizations for both low-level arithmetic and high-level operations, including minimizing instructions for modular operations, enhancing hardware utilization for base conversion, implementing efficient reuse strategies, and introducing intra-arithmetic and inner-conversion fusion methods, thus decreasing the overall computational and memory consumption.
Leveraging our framework, we offer comprehensive comparative analyses. Our performance evaluation showcases a marked speed improvement, achieving 31.9× over OpenFHE running on a multi-threaded CPU and 39.7% and 29.9% improvement, respectively, over the state-of-the-art GPU BEHZ implementation. Our implementation of the leveled HPS variant records up to 4× speedup over other variants, positioning it as a highly promising alternative for specific applications.
In this work, we present an in-depth study focusing on accelerating and comparing BFV variants on GPUs, including Bajard-Eynard-Hasan-Zucca (BEHZ), Halevi-Polyakov-Shoup (HPS), and other recent variants. We introduce a universal framework accommodating all variants, propose optimized BEHZ implementation, and first support HPS variants with large parameter sets on GPUs. Moreover, we devise several optimizations for both low-level arithmetic and high-level operations, including minimizing instructions for modular operations, enhancing hardware utilization for base conversion, implementing efficient reuse strategies, and introducing intra-arithmetic and inner-conversion fusion methods, thus decreasing the overall computational and memory consumption.
Leveraging our framework, we offer comprehensive comparative analyses. Our performance evaluation showcases a marked speed improvement, achieving 31.9× over OpenFHE running on a multi-threaded CPU and 39.7% and 29.9% improvement, respectively, over the state-of-the-art GPU BEHZ implementation. Our implementation of the leveled HPS variant records up to 4× speedup over other variants, positioning it as a highly promising alternative for specific applications.
Additional news items may be found on the IACR news page.