Abstract
Fully Homomorphic Encryption (FHE) enables encrypted data processing on untrusted cloud servers, crucial for privacy-sensitive applications. Despite its potential, performance overheads (about $10,000\times$ slower) limit adoption. ASIC accelerators outperform GPUs/FPGAs by optimizing specific operations but rely on costly 7nm processes and large on-chip memory, hindering cost-effective deployment. Balancing efficiency with manufacturing constraints remains critical. This paper presents TensorFHE+, a GPU-optimized FHE acceleration framework leveraging Tensor Cores to accelerate Number Theoretic Transform (NTT) operations. Key innovations include: 1) Decomposing CKKS kernels into vector/matrix operations for hardware utilization; 2) Vectorized modulo arithmetic; 3) Data layout optimization for memory efficiency. Evaluated on NVIDIA A100, TensorFHE+ outperforms TensorFHE[1] by $1.44\times$ in average (up to $1.69\times$ on ResNet-20) and surpasses prior GPU implementations [2], [3]. The design also demonstrates compatibility with commercial linear algebra accelerators, enabling efficient FHE deployment.
Funding Information
  • National Key R&D Program of China (2023YFB3106200)
  • National Natural Science Foundation of China (62572020)
  • National Natural Science Foundation of China (T2425023)
  • National Natural Science Foundation of China (624B2016)
  • National Natural Science Foundation of China (62202028)
  • National Natural Science Foundation of China (62172025)
  • National Natural Science Foundation of China (U2241213)