Does deepep-tests scripts support nsys profile?

it uses kineto to capture kernel durations. does it still work along with `nsys profile -t cuda`?

if it doesn't, can we add a command line option or some env variable to by-pass kineto data capturing and calculation, so `nsys profile` can work?

Start with the deepep-tests scripts and run them under `nsys profile -t cuda` to determine whether kineto capture and calculation interfere with profiling. Done means confirming compatibility or documenting and exposing a command-line option or environment variable to bypass kineto data capture and calculation.

Written by the indexing model from the issue text.