NVIDIA open-source le collectif Reshard pour accélérer le refit de poids RL jusqu’à 5,7×
![[object Object]](https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/green-cube.webp)
NCCL M2N Reshard ne déplace que les shards en intersection ; NeMo RL affiche 2,7–5,7× vs gather-and-broadcast.
Pourquoi c’est important
NVIDIA a décrit Reshard, un collectif qui traite le refit RL comme un transfert DTensor-à-DTensor.
Faits clés
- 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