Skip to content

Simple continous power capping implementation #123

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
Open
wants to merge 4 commits into
base: dev
Choose a base branch
from
Open

Simple continous power capping implementation #123

wants to merge 4 commits into from

Conversation

huwzpf
Copy link
Owner

@huwzpf huwzpf commented Feb 9, 2025 •
edited
Loading

Changes in this PR:

  • Fixed reading energy used in python scripts
  • Added slave handling for __cudampi__CONFIGUREPOWERCAP
  • Added option to choose from 2 power capping strategies
  • With continous strategy:
    • Acceptable power cap ranges are read during initialization and power caps for each device are initialized as equal split of power budget, with each device receiveing no less than 0.33 of it's acceptable power cap range
    • During synchronization, first devices are selected like previously, but using mininum power cap as power consumption. After devices are selected, remaining power is split equally.

@huwzpf huwzpf force-pushed the continous-power-capping branch from 451214c to ae72778 Compare February 9, 2025 23:20
@huwzpf
Copy link
Owner Author

huwzpf commented Feb 9, 2025

Opens:

  • Should we also consider master's CPU ?
  • How to approach including CPU power usage in GPU power usage ?
  • Should power cap redistribution be done only once with every synchronize ?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@huwzpf