cs.DCNov 13, 2026

Compiling Specialized GPU-Initiated Networking

Authors: Javid Baydamirli, Ismayil Ismayilov, Kaan Oktay, Didem Unat

Organizations: Koc¸ University, Turkey · fal

Abstract

Machine learning systems specialize computation to the workload and hardware, but communication remains in generic libraries whose device code and transport decisions are not visible to high-level compilers. We present Obaleth, a compiler that makes GPU-initiated communication part of a kernel's intermediate representation, co-optimizes it alongside computation, and lowers it to specialized transport code. Obaleth plans completions around buffer lifetimes and optimizes memory ordering, synchronization, and batching across NVIDIA and AMD GPUs. On H100 GPUs, Obaleth sends small messages up to 3.3×\times faster, and on 256 GPUs runs all-to-all 1.25×\times and 2.3×\times faster than NVSHMEM and NCCL GIN, with much smaller memory and register use. In vLLM, it serves Qwen3-235B on 16 B200 GPUs and answers requests up to 1.29×\times faster than with DeepEP.

Figures & tables

Appendix figures & tables9 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. HetCCL: Efficient LLM Training on Heterogeneous Vendor GPUs

    Jan 30, 2026Heehoon Kim, Chris Jaehwan Lee, Taejeoung Kim +6Heterogeneous ComputingDistributed Training

  2. UCCL-Zip: Lossless Compression Supercharged GPU Communication

    Apr 19, 2026Shuang Ma, Chon Lam Lao, Zhiying Xu +8GPU Kernel OptimizationGPU Acceleration