PCI Express
计算机科学
延迟(音频)
调度(生产过程)
利用
关键路径法
低延迟(资本市场)
嵌入式系统
架空(工程)
软件部署
并行计算
推论
计算机网络
地铁列车时刻表
操作系统
分布式计算
实时计算
数据传输
现场可编程门阵列
向后兼容性
偏移量(计算机科学)
数字信号处理
作者
Yao Fei,Jin Fang,Size Zheng,Gongming Zhao,Hongli Xu,Jiacheng Zhu,Zhijing Xin
标识
DOI:10.48550/arxiv.2609.34340
摘要
AlltoAllv communication is a critical primitive in distributed large-model inference, particularly for mixture-of-experts (MoE) models. The growing adoption of PCIe GPU systems for cost-efficient inference makes AlltoAllv performance on these systems increasingly important. Without a dedicated scale-up interconnect (e.g., NVLink or Infinity Fabric), PCIe GPU systems carry both intra-node and inter-node traffic through the PCIe hierarchy, where concurrent transfers can contend for PCIe link bandwidth. This link contention, compounded by skewed traffic distributions and dynamic traffic demand, makes efficient AlltoAllv scheduling challenging. Existing approaches are either poorly suited to PCIe GPU systems or incur substantial schedule synthesis overhead that reduces their practicality in real-world deployments. We present VarioPath, an efficient AlltoAllv scheduling framework for PCIe GPU systems. It combines an offline topology-aware analyzer with an online demand-aware scheduler. The analyzer records contention-free transfer patterns as AlltoAllv channels and exploits topology symmetry to build a compact catalog for efficient search. The online scheduler decomposes each AlltoAllv invocation's demand across a sequence of channels, adapting to rapidly changing and skewed traffic while incurring low planning overhead. Evaluation on four platforms (up to 256 GPUs) shows average AlltoAllv speedups of 5.88x over FAST and 1.72x over DeepEP. End-to-end experiments show that VarioPath reduces Qwen3 inference latency by up to 27.2% and Wan2.1 generation latency by 6.1%.
科研通智能强力驱动
Strongly Powered by AbleSci AI