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

Avoid poisoning process with CUDA calls as soon as importing #6810

Merged
merged 6 commits into from
Dec 12, 2024
Merged

Avoid poisoning process with CUDA calls as soon as importing #6810

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
0e67208
Avoid poisoning process with CUDA calls as soon as importing
HollowMan6 Nov 29, 2024
907dedd
Merge branch 'master' into pr
loadams Dec 10, 2024
cc762aa
Merge branch 'master' into pr
loadams Dec 10, 2024
1dc0be3
Merge branch 'master' into pr
loadams Dec 10, 2024
de0f21d
Merge branch 'master' into pr
loadams Dec 11, 2024
5197e35
Merge branch 'master' into pr
loadams Dec 12, 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
7 changes: 6 additions & 1 deletion accelerator/real_accelerator.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -167,7 +167,12 @@ def get_accelerator():
import torch

# Determine if we are on a GPU or x86 CPU with torch.
if torch.cuda.is_available(): #ignore-cuda
# "torch.cuda.is_available()" provides a stronger guarantee, #ignore-cuda
# ensuring that we are free from CUDA initialization errors.
# While "torch.cuda.device_count() > 0" check ensures that #ignore-cuda
# we won't try to do any CUDA calls when no device is available
# For reference: https://github.com/microsoft/DeepSpeed/pull/6810
if torch.cuda.device_count() > 0 and torch.cuda.is_available(): #ignore-cuda
accelerator_name = "cuda"
else:
if accel_logger is not None:
Expand Down