compute_ru

User Tools

Site Tools

Trace: • pythongpu • gpupython • begin
compute:pythongpu

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
compute:pythongpu [2024/12/11 15:42] – bjornscompute:pythongpu [2024/12/11 15:48] (current) – bjorns
Line 16: Line 16:
 ''if torch.cuda.is_available() print("GPU")'' ''if torch.cuda.is_available() print("GPU")''
 \\ \\
-But you must run this code test with sbatch\\ +But you must run this code test with **sbatch**, send your job to the cluster\\ 
-Create a python file dummy.py with these command\\+Create a python file dummy.py with the test-GPU command\\
 \\ \\
 Update the sbatch script file, mine is called **oneJob.sh**\\ Update the sbatch script file, mine is called **oneJob.sh**\\
Line 29: Line 29:
 Then check your log file (default name is myJob.log) for the result\\ Then check your log file (default name is myJob.log) for the result\\
 \\ \\
-Installing Tensorflow is same process, we recommend a different virtual environment if you want to try both of them\\+Installing Tensorflow is same process, we recommend using a different virtual environment if you want to try both of them\\
 Make sure to instruct pip in this way, to get all necessary packages ''pip install "tensorflow[and-cuda]"''\\ Make sure to instruct pip in this way, to get all necessary packages ''pip install "tensorflow[and-cuda]"''\\
 \\ \\
compute/pythongpu.1733931726.txt.gz · Last modified: 2024/12/11 15:42 by bjorns