Profiling across multiple mpi machines

[nikhilbhendawade]

I was wondering how to profile deep learning model across multiple nodes (mpi machines). Simply using “nsys profile …” doesn’t work.