You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The intra-node p2p tput using the copy engine is still low, compared to NCCL/RCCL's SM copy. To reproduce: torchrun --nproc_per_node=8 benchmark_uccl_collective.py --ring. Is this inherent to the copy engine?