Skip to content

Latest commit

 

History

History
40 lines (31 loc) · 2.05 KB

README.md

File metadata and controls

40 lines (31 loc) · 2.05 KB

HyLink

A two-layer collective communication framework that unlocks underutilized PCIe bandwidth and jointly harnesses dedicated interconnects and the PCIe bus.

Overview.

Features

  • 💡 PCIe CCL accelerates collective communication on the PCIe path by making the CPU-rooted PCIe hierarchy explicit and treating NUMA-local host memory as an active routing point, powered by a CPU-driven DSL for topology-aware schedules and an auto-pipelined runtime built on GPU copy engines.
  • 🔀 Multipath Orchestrator splits each collective across the dedicated link (via the vendor CCL) and the PCIe path (via PCIe CCL) on independent streams with a fork–join model, and a feedback-driven adaptive balancer retunes the split ratio at runtime to keep the two paths balanced under PCIe interference.

Usage

See each component's README for build and usage instructions:

Cite

Your citations are greatly appreciated. 🥰

  • ACM Reference

    Yuezheng Liu, Menghao Zhang, Xuebin Song, Juner Shen, Chunming Hu, and Xudong Liu. 2026. Towards High-Throughput Collective Communication via Harnessing PCIe and Dedicated Interconnects. In Proceedings of the 2026 ACM SIGOPS Annual Technical Conference (ATC 2026).

  • Bibtex
    @inproceedings{hylink2026,
        author    = {Liu, Yuezheng and Zhang, Menghao and Song, Xuebin and Shen, Juner and Hu, Chunming and Liu, Xudong},
        title     = {Towards High-Throughput Collective Communication via Harnessing PCIe and Dedicated Interconnects},
        year      = {2026},
        booktitle = {Proceedings of the 2026 ACM SIGOPS Annual Technical Conference},
        series    = {ATC 2026}
    }