Skip to content

[ROCm] xdist: select GPU from runner-allocated ROCR slice #793

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
mminutoli wants to merge 1 commit into
base: rocm-jaxlib-v0.9.1
Choose a base branch
from
Open

[ROCm] xdist: select GPU from runner-allocated ROCR slice #793

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
12 changes: 10 additions & 2 deletions conftest.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,9 +79,17 @@ def pytest_collection() -> None:
if not xdist_worker_name.startswith("gw"):
return
xdist_worker_number = int(xdist_worker_name[len("gw") :])
os.environ["ROCR_VISIBLE_DEVICES"] = str(
xdist_worker_number % num_rocm_devices
allocated = os.environ.get("ROCR_VISIBLE_DEVICES")
allocated_tokens = (
[t.strip() for t in allocated.split(",") if t.strip()]
if allocated
else []
)
if allocated_tokens:
selected = allocated_tokens[xdist_worker_number % len(allocated_tokens)]
else:
selected = str(xdist_worker_number % num_rocm_devices)
os.environ["ROCR_VISIBLE_DEVICES"] = selected
# ROCR_VISIBLE_DEVICES filters HSA to a single physical device, which
# becomes HIP index 0. The container env-file may preset
# HIP_VISIBLE_DEVICES to all GPUs; override to "0" so HIP doesn't try to
Expand Down