NVIDIAがReshardコレクティブをオープンソース化、RL重みrefitを最大5.7倍高速化
![[object Object]](https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/green-cube.webp)
NCCL M2N Reshardは交差シャードのみ転送；NeMo RLでgather-and-broadcast比2.7–5.7倍。
なぜ重要か
NVIDIAはReshardを説明：RL重みrefitをDTensor間転送とし、理論最小の交差シャードだけを動かす。
要点
- New Reshard collective for DTensor-to-DTensor refit
- Open NCCL M2N Reshard reference implementation
- 2.7×–5.7× faster than gather-and-broadcast in NeMo RL
- Benchmarked on DGX H100 and GB200
- Hybrid path: FFN projections via Reshard; small weights legacy
- Aimed at trillion-parameter RL post-training payloads