Skip to content

fix: synchronize before the pre-broadcast empty_cache #3383

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

Merged
mikasenghaas merged 1 commit into from
Aug 25, 2026
Merged

fix: synchronize before the pre-broadcast empty_cache #3383

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
7 changes: 6 additions & 1 deletion src/prime_rl/trainer/rl/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -604,7 +604,12 @@ def train(config: TrainerConfig):
broadcast_weights_start_time = time.perf_counter()
# The per-layer gather + fp8 conversion peaks ~50 GiB above the
# resident weights; release cached blocks (incl. offload-stream
# pools) so the broadcast gets the full headroom.
# pools) so the broadcast gets the full headroom. Drain all
# pending work first: empty_cache returns blocks to the driver,
# so a still-running kernel holding a cached block (e.g. the
# optimizer step's tail) faults with an illegal memory access
# once its block is freed under it.
torch.cuda.synchronize()
torch.cuda.empty_cache()
weight_sender.broadcast(model, step=progress.step)
broadcast_weights_time = time.perf_counter() - broadcast_weights_start_time
Expand Down