GPU Acceleration of TFHE-Based High-Precision Nonlinear Layers for Encrypted LLM Inference
Peking University（北京大学）；Mohamed bin Zayed University of Artificial Intelligence（穆罕默德·本·扎耶德人工智能大学）；Xiamen University（厦门大学）；Nanjing University（南京大学）；Ant Group（蚂蚁集团）
Guoci Chen, Xiurui Pan, Qiao Li, Bo Mao, Congming Gao, Chengying Huan, Mingzhe Zhang, Jie Zhang
Deploying large language models (LLMs) as cloud services raises privacy concerns as inference may leak sensitive data. Fully Homomorphic Encryption (FHE) allows computation on encrypted data, but current FHE methods struggle with efficient and precise nonlinear function evaluation. Specifically, CKKS-based approaches require high-degree polynomial approximations, which are costly when target precision increases. Alternatively, TFHE's Programmable Bootstrapping (PBS) outperforms CKKS by offering exact lookup-table evaluation. But it lacks high-precision implementations of LLM nonlinear layers and underutilizes GPU resources.
We propose \emph{TIGER}, the first GPU-accelerated framework for high-precision TFHE-based nonlinear LLM layer evaluation. TIGER offers: (1) GPU-optimized WoP-PBS method combined with numerical algorithms to surpass native lookup-table precision limits on nonlinear functions; (2) high-precision and efficient implementations of key nonlinear layers, enabling practical encrypted inference; (3) batch-driven design exploiting inter-input parallelism to boost GPU efficiency. TIGER achieves 7.17$\times$, 16.68$\times$, and 17.05$\times$ speedups over a CPU baseline for GELU, Softmax, and LayerNorm, respectively.
Efficient Encrypted Computation in Convolutional Spiking Neural Networks with TFHE
School of Cyber Science and Engineering, Huazhong University of Science and Technology（华中科技大学网络空间安全学院）；School of Mathematics and Statistics, Huazhong University of Science and Technology（华中科技大学数学与统计学院）；Center for Mathematical Science, Huazhong University of Science and Technology（华中科技大学数学中心）；Steklov-Wuhan Institute for Mathematical Exploration, Huazhong University of Science and Technology（华中科技大学斯捷克洛夫-武汉数学探索研究所）；Department of Mathematics and Department of Physics, Great Bay University（大湾区大学数学与物理学院）；Guangdong Provincial Key Laboratory of Mathematical and Neural Dynamical Systems（广东省数学与神经动力系统重点实验室）
Scaling up FHE-based Privacy-Preserving ML: Higher Throughput, Longer Inputs for LLama-3-8B
Graduate School of Convergence Science and Technology, Seoul National University（首尔大学融合科学技术研究生院）；CryptoLab, Inc.（CryptoLab公司）；Department of Mathematics, Seoul National University（首尔大学数学系）
