Metraipolis › archive › dc_62bb007a595f

VarioPath: Workload-Aware All-to-All Communication for PCIe GPU Clusters — arXiv:2609.34340

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.34340
License (as archived)
CC0-1.0
Archived
2026-10-07T22:12:11.953Z
SHA-256 of the archived text
4a71240b578204784de8885938330277c99435594c677704e82659b90f4ab17e
Stable URL of this page
https://metraipolis.com/documento/dc_62bb007a595f

How to cite

VarioPath: Workload-Aware All-to-All Communication for PCIe GPU Clusters — arXiv:2609.34340 — arXiv (en), https://arxiv.org/abs/2609.34340 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_62bb007a595f (Metraipolis archive dc_62bb007a595f, sha256 4a71240b578204784de8885938330277c99435594c677704e82659b90f4ab17e)
BibTeX
@misc{dc62bb007a595f,
  title = {VarioPath: Workload-Aware All-to-All Communication for PCIe GPU Clusters — arXiv:2609.34340},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.34340},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_62bb007a595f; sha256 4a71240b578204784de8885938330277c99435594c677704e82659b90f4ab17e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34340. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

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%.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-29 in cs.DC).
Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC) ; Networking and Internet Architecture (cs.NI).
Pagina del lavoro: https://arxiv.org/abs/2609.34340 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T21:53:13.708Z dalla pagina https://arxiv.org/catchup/cs.DC/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).