Skip to content

Explicitly delete objects in async checkpoint worker loop and call GC #2740

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
apaithankar wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Explicitly delete objects in async checkpoint worker loop and call GC #2740

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3d444b4
explicitly delete objects in async ckpt worker loop and call gc
apaithankar Dec 23, 2025
f06f43c
Merge branch 'main' into gc_changes
apaithankar Dec 27, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions megatron/core/dist_checkpointing/strategies/async_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -436,7 +436,6 @@ def __del__(self):
self.close()

@staticmethod
@_disable_gc()
def async_loop(
rank: int,
queue: mp.JoinableQueue,
Expand Down Expand Up @@ -486,7 +485,9 @@ def async_loop(
logger.debug(f"{rank} has completed saving {item.call_idx}")
comp_q.put(item.call_idx)
queue.task_done()

del async_fn_args
del item
gc.collect()
logger.info(f"PersistentAsyncCaller: persistent ckpt worker for {rank} has terminated")


Expand Down
Loading