Add Nsight Docs

[anshgupta1234]

This PR adds --nsys and --ncu to the documentation. I wasn't sure how much detail to add, please let me know if there's anything important that I missed :)

[sbryngelson]

why did you change the CI file

[anshgupta1234]

why did you change the CI file

My apologies, I reverted it.

[sbryngelson]

Please add links to what nsight compute and nsight sys are and their official documentation, and how to use additional flags for them. You also should describe what they are used for.

[anshgupta1234]

@sbryngelson Sounds good, I'll add that.

[henryleberre]

@sbryngelson @anshgupta1234 I pushed to this PR the feature that lets you specify additional arguments to nsys and ncu. For brevity, I won't explain why but it requires that --nsys (or --ncu) be specified as the last command line argument. You can just append to the invocation the arguments you want to have forwarded.

For example you could:

$ ./mfc.sh run case.py [(optional) case file arguments] -N 5 -n 2 --gpu --no-mpi -j $(nproc) -t simulation -e batch .... --nsys [nsys arguments here]

This isn't necessarily realistic but it shows how you can use all the features together.

[sbryngelson]

@henryleberre thanks! should be useful. can you add this to the documentation as well, @anshgupta1234 ?