NVIDIA releases NCCL 2.28
Released NCCL 2.28 — a Device API that lets kernels initiate communication directly, GPU-Initiated Networking, Copy Engine collectives that use no SMs, and symmetric memory support
Why it matters
A release that supports fusing communication and compute at the kernel level, changing how the communication bottleneck in large-scale distributed training and inference is tackled
Source — Primary sources first; secondary reports are labelled “Press”
Officialdeveloper.nvidia.comhttps://developer.nvidia.com/blog/fusing-communication-and-compute-with-new-device-api-and-copy-engine-collectives-in-nvidia-nccl-2-28/Added in Week 35, 2026 · See that week