-
Notifications
You must be signed in to change notification settings - Fork 26
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.
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
Move GPU Enabling to a subscript #7
Comments
This is now a function defined in metashape_pipeline_functions.py. It can be included/excluded as a call from a control script as a means of enabling/not enabling. I have not yet implemented GPU_num because I could not immediately see how to specify the number. Do you feel the "GPU subscript" component of this issue is addressed? If so, we can close this and create a separate ticket for the GPU_num issue. |
We do need to modify this if we plan to ever use multiple GPUs.
A small function to translate from an optional number, default=1 to the appropriate bitmask will do. |
Will this require each parallel instance of the script to know which specific GPUs the other instances are using? |
No, since the resource sets the number of GPUs available they will always be labeled 1 to n. So typically we would use all available, so the user can just provide n (or slurm as an env variable might) and we translate to enable all for however many n that is. |
Add derived config creation R function
The code to check for GPUs, report their names and enable if available or configured can be it's own script.
Configuration should include
GPU_enable = True/False
GPU_num = numeric (number of GPUs to use if available)
The text was updated successfully, but these errors were encountered: