Skip to content
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

Jump to bottom

feat(pt): support CPU parallel training with PT #4224

Merged
merged 3 commits into from
Oct 23, 2024
Merged

feat(pt): support CPU parallel training with PT #4224

Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
6acfd3b
feat(pt): support CPU parallel training with PT
iProzd Oct 16, 2024
0b0d943
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Oct 16, 2024
f9b43e8
Update main.py
iProzd Oct 23, 2024
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
14 changes: 12 additions & 2 deletions deepmd/pt/entrypoints/main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,8 +105,18 @@
local_rank = os.environ.get("LOCAL_RANK")
if local_rank is not None:
local_rank = int(local_rank)
assert dist.is_nccl_available()
dist.init_process_group(backend="nccl")
nccl_available = dist.is_nccl_available()
gloo_available = dist.is_gloo_available()

Check warning on line 109 in deepmd/pt/entrypoints/main.py

View check run for this annotation

Codecov / codecov/patch

deepmd/pt/entrypoints/main.py#L108-L109 

Added lines #L108 - L109 were not covered by tests
# nccl first
if nccl_available:
backend = "nccl"
elif gloo_available:
backend = "gloo"

Check warning on line 114 in deepmd/pt/entrypoints/main.py

View check run for this annotation

Codecov / codecov/patch

deepmd/pt/entrypoints/main.py#L111-L114 

Added lines #L111 - L114 were not covered by tests
iProzd marked this conversation as resolved.
Show resolved Hide resolved
iProzd marked this conversation as resolved.
Show resolved Hide resolved
else:
raise RuntimeError(

Check warning on line 116 in deepmd/pt/entrypoints/main.py

View check run for this annotation

Codecov / codecov/patch

deepmd/pt/entrypoints/main.py#L116 

Added line #L116 was not covered by tests
"No suitable backend found. Neither NCCL nor Gloo is available."
)
dist.init_process_group(backend=backend)

Check warning on line 119 in deepmd/pt/entrypoints/main.py

View check run for this annotation

Codecov / codecov/patch

deepmd/pt/entrypoints/main.py#L119 

Added line #L119 was not covered by tests
iProzd marked this conversation as resolved.
Show resolved Hide resolved

def prepare_trainer_input_single(
model_params_single, data_dict_single, rank=0, seed=None
Expand Down