Question about Nsight Compute's application range replay support data collection for multi-node, multi-GPU setups under NCCL

Developer Tools Nsight Compute
1

Hi,
I would like to know if the latest version of Nsight Compute supports data collection across multi-node multi-GPU configurations?
For example, I have two nodes, each with 2 GPUs.
How can I use NCU to collect data from all 4 GPUs?
Is there an official guide or manual that explains this?
Thanks.

2

Hi, @FlyK

Please refer Nsight Compute CLI :: Nsight Compute Documentation for applications using MPI.

To profile multi-node submissions, one instance of NVIDIA Nsight Compute CLI can be used per node. Ensure that you specify unique report files per rank.
mpirun [mpi arguments] ncu -o report_%q{OMPI_COMM_WORLD_RANK} [app arguments]

3

Thank you,
Best,

4

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.