Add an API to DDP for dynamically updating the underlying process group.

[pritamdamania87]

Motivation

If we would like to reinitialize DDP with a different PG with torch.compile, we need to do the following:

del old_ddp
del old_pg
pg = init_pg(...)
ddp = DDP(pg)
model = torch.compile(DDP)

This results in recompilation of the entire model and is very expensive. Since the only thing we need to update is the PG, we should be able to do this without having to compile the model again.

Proposal

As a result, in this PR I've introduced an _update_process_group API which can dynamically update the underlying ProcessGroup used by DDP without needing to reinitialize DDP again.

cc @mrshenli @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @osalpekar @jiayisuse @H-Huang @kwen2501 @awgu @penguinwu @fegin @XilunWu @wanchaol @fduwjj @wz337 @tianyu-l @wconstab @yf225 @kiukchung @d4l3k @LucasLLC

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/113580

• 📄 Preview Python docs built from this PR
• 📄 Preview C++ docs built from this PR
• ❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (1 Unrelated Failure)

As of commit 3c754c4 with merge base a2c32b8 ():

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• slow / linux-focal-rocm5.6-py3.8 / test (slow, 1, 1, linux.rocm.gpu) (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[XilunWu]

this PR has a lot of re-formatting which I suggest we exclude.

[pritamdamania87]

this PR has a lot of re-formatting which I suggest we exclude.

@XilunWu This came from lintrunner -a automatically. Not sure if lint will fail if I don't apply those changes.

[wconstab]

this PR has a lot of re-formatting which I suggest we exclude.

@XilunWu This came from lintrunner -a automatically. Not sure if lint will fail if I don't apply those changes.

actually while i hate to be pedantic i think in this case since there are so many lines of code affected, it'd be worth splitting the PR into a stack where the first PR just applies lintrunner and the second PR just has your 'real' changes.

[wconstab]

@rohan-varma @fegin any issue with this PR? (I can do a more detailed review if you dont want to but any objection on the addition of these APIs in principle?)

[pritamdamania87]

actually while i hate to be pedantic i think in this case since there are so many lines of code affected, it'd be worth splitting the PR into a stack where the first PR just applies lintrunner and the second PR just has your 'real' changes.

So I had to run lintrunner -a since make lint was failing locally for me. If I remove the formatting changes won't lint fail on this PR and as a result block merge?

[wconstab]

actually while i hate to be pedantic i think in this case since there are so many lines of code affected, it'd be worth splitting the PR into a stack where the first PR just applies lintrunner and the second PR just has your 'real' changes.

So I had to run lintrunner -a since make lint was failing locally for me. If I remove the formatting changes won't lint fail on this PR and as a result block merge?

my suggestion was to create a 'stack' where your first pr does the lint changes and then your second PR which is 'on top' does the changes. (are you familiar with ghstack? pip install ghstack and instructions on its github page)

[pritamdamania87]

my suggestion was to create a 'stack' where your first pr does the lint changes and then your second PR which is 'on top' does the changes. (are you familiar with ghstack? pip install ghstack and instructions on its github page)

Looks like maybe there was something broken in my local setup causing the linter to change files. Removing the formatting still passes CI though

[fduwjj]

Overall, this looks good to me. Unblock for now but one thing I am not sure is that, are the changes in the reducer the only places we need to change? tbh, I am not reading reducer's code enough. Maybe, can @rohan-varma or @fegin also take a look? I also added more CI lables.

[pritamdamania87]

@pytorchbot merge

[pytorchmergebot]

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

[fegin]

Forgot to leave the comments. The PR looks good to me.