add profiling to ProcessGroupBaby

[d4l3k]

Currently ProcessGroupBaby doesn't support any profiling as the record_function pieces will run in the subprocess.

We either need to figure out some way to forward those profiling information from the child process -- or we can just add RecordFunction support to _run_func and _BabyWork.

We probably want to be able to track the async completion (i.e. via get_future) but that may be pretty expensive and also isn't super accurate since completion via get_future immediately completes for NCCL

It may be better to track blocking time (i.e. _run_func/wait) as well

Relevant code:

• record_function in PyTorch https://github.com/pytorch/pytorch/blob/b1a81a4a650a17988ef9289b1a7697a384c48b26/torch/autograd/profiler.py#L703
• _run_func https://github.com/pytorch/torchft/blob/main/torchft/process_group.py#L1187-L1223

[ekdnam]

I would like to work on this!