用统一编译框架加速安全大模型推理，性能提升超24%。
FuseFSS: Efficient Secure LLM Inference with Function Secret Sharing
- 将每个算子的专用协议替换为统一编译管道，简化实现。
- 在BERT和GPT模型上实现1.24到1.50倍端到端提速，通信量降9%~16%。
- 适合需要高效安全推理的云端大模型应用，尤其关注隐私保护场景。
两服务器安全推理使客户端在不泄露提示词或嵌入向量的情况下查询托管的大语言模型（LLM）。基于函数秘密共享（FSS）的现有GPU系统虽使线性层高效，但定点非线性运算和辅助操作仍是瓶颈，因每个算子通常需定制协议，包含独立比较、进位修正和预处理数据。本文提出FuseFSS，一个编译器，将每算子协议设计转化为单一编译流水线。对每个标量定点算子，紧凑规格列出其区间划分、低次多项式部分及所需判定比特。编译器生成两次批处理的FSS评估：一次打包比较返回所有判定比特，一次向量区间查找返回激活系数与常数。相比当前最先进的基于FSS的GPU安全推理，FuseFSS保持精度的同时，在BERT和GPT类模型上实现1.24×–1.50×端到端加速，在线通信减少9%–16%；预处理也更轻量，密钥生成时间降低14%–23%，密钥大小减小20%–24%。
原文摘要 · Abstract (English)
Two-server secure inference allows a client to query a hosted large language model (LLM) without revealing prompts or embeddings. Recent GPU systems based on function secret sharing (FSS) make linear layers efficient, but fixed-point nonlinearities and helper operations remain a bottleneck because each operator is typically implemented as a bespoke protocol with its own comparisons, wrap-around corrections, and preprocessing material. We present FuseFSS, a compiler that replaces per-operator protocol design with a single compilation pipeline. For each scalar fixed-point operator, a compact specification lists its interval partition, low-degree arithmetic pieces, and required predicate bits. The compiler emits two batched FSS evaluations on the public masked value: one packed comparison that returns all predicate bits, and one vector interval lookup that returns the active coefficients and constants. Compared to the current state-of-the-art FSS-based GPU secure inference, FuseFSS preserves accuracy while achieving a $1.24\times$--$1.50\times$ end-to-end speedup and reducing online communication by $9\%$--$16\%$ on BERT and GPT-style models; preprocessing is also lighter, with $14\%$--$23\%$ lower key-generation time and $20\%$--$24\%$ smaller keys.
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。