利用稀疏性加速加密神经网络矩阵乘，GPU性能提升三倍。
GPU Acceleration of Sparse Fully Homomorphic Encrypted DNNs
- 针对加密神经网络的矩阵乘，利用双操作数稀疏性优化
- 在AMD GPU上实现，运行时间比CPU快3.0倍，复杂度从立方降至半线性
- 适合需要隐私保护的深度学习部署场景
全同态加密（FHE）作为密码学原语与系统挑战备受关注。借助加速计算最新进展，FHE在机器学习与信息安全领域展现巨大潜力。本文聚焦深度神经网络中最耗时的操作——矩阵乘法（matmul），从硬件角度出发，将其适配于AMD GPU执行。通过采用专为GPU设计的开源FHE库FIDESlib，提出一种新优化方法，利用两个操作数的稀疏性，显著提升密文矩阵乘的运行效率。实验表明，该稀疏矩阵乘实现比其CPU版本最高快3.0倍，时间复杂度从立方级降至半线性，优于现有FHE矩阵乘方案。
原文摘要 · Abstract (English)
Fully homomorphic encryption (FHE) has recently attracted significant attention as both a cryptographic primitive and a systems challenge. Given the latest advances in accelerated computing, FHE presents a promising opportunity for progress, with applications ranging from machine learning to information security. We target the most computationally intensive operation in deep neural networks from a hardware perspective, matrix multiplication (matmul), and adapt it for execution on AMD GPUs. We propose a new optimized method that improves the runtime and complexity of ciphertext matmul by using FIDESlib, a recent open-source FHE library designed specifically for GPUs. By exploiting sparsity in both operands, our sparse matmul implementation outperforms its CPU counterpart by up to $3.0\times$ and reduces the time complexity from cubic to semi-linear, demonstrating an improvement over existing FHE matmul implementations.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。