Yang Zhou, University of California, Berkeley and University of California, Davis; Zhongjie Chen, Tsinghua University; Ziming Mao, University of California, Berkeley; ChonLam Lao, Harvard University; Shuo Yang, University of California, Berkeley; Pravein Govindan Kannan, IBM Research; Xizhi Zhang, Tsinghua University; Jiaqi Gao, Independent Researcher; Yilong Zhao and Yongji Wu, University of California, Berkeley; Kaichao You, Independent Researcher; Fengyuan Ren, Tsinghua University; Zhiying Xu, Amazon Web Services; Costin Raiciu, Broadcom and University Pollitehnica of Bucharest; Ion Stoica, University of California, Berkeley
Fast-evolving machine learning (ML) workloads have increasing requirements for networking. However, host network transport on RDMA NICs is hard to evolve, causing problems for ML workloads. For example, single-path RDMA traffic is prone to flow collisions that severely degrade collective communication performance. We present UCCL-Tran, an extensible software transport layer to evolve GPU networking. UCCL-Tran decouples the data path and control path of existing RDMA NICs and efficiently runs the control-path transport on host CPUs. This software extensibility brings in transport innovations that cannot be achieved in hardware for ML workloads, e.g., a multipath transport to resolve flow collisions. ML collectives atop UCCL-Tran achieve up to 4.5× higher performance compared to existing RDMA NICs.
OSDI '26 Open Access Sponsored by
King Abdullah University of Science and Technology (KAUST)
Open Access Media
USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.
author = {Yang Zhou and Zhongjie Chen and Ziming Mao and ChonLam Lao and Shuo Yang and Pravein Govindan Kannan and Xizhi Zhang and Jiaqi Gao and Yilong Zhao and Yongji Wu and Kaichao You and Fengyuan Ren and Zhiying Xu and Costin Raiciu and Ion Stoica},
title = {{UCCL-Tran}: An Extensible Software Transport Layer for {GPU} Networking},
booktitle = {20th USENIX Symposium on Operating Systems Design and Implementation (OSDI 26)},
year = {2026},
isbn = {978-1-939133-55-7},
address = {Seattle, WA},
pages = {1143--1166},
url = {https://www.usenix.org/conference/osdi26/presentation/zhou-yang},
publisher = {USENIX Association},
month = jul
}
