NVIDIA Ships NCCL 2.31.2 with GPU-Driven Communication and 0-SM Collectives

NVIDIA released NCCL 2.31.2 with GPU-driven CFT/RMA fine-grained communication, per-collective tuning, NVLS+PAT improvements, 0-SM collectives, and multi-NIC optimizations, targeting Blackwell-scale training.

2026-09-23 ~ 2026-09-23 · 2 related posts

• NCCL 2.31.2 ships GPU-driven CFT, per-collective tuning and 0-SM collectives for Blackwell-scale training — SkyLi0n · 2026-09-23
• NCCL 2.31.2 ships GPU-driven CFT/RMA, 0-SM collectives, better multi-NIC for Blackwell-scale training — SkyLi0n · 2026-09-23