mila-iqia / milatools

Tools to connect to and interact with the Mila cluster
MIT License
62 stars 12 forks source link

[v0.1.3-post.1+db9b7bc] Issue running the command `mila code` #113

Closed shrutij01 closed 6 months ago

shrutij01 commented 6 months ago

Make sure you can reproduce the issue with the latest version available

I am on version milatools v0.1.3-post.1+db9b7b (the latest version)

What command did you run?

mila code $HOME_DIR --alloc --gres=gpu:1 -c 4 --time=2:00:00 --partition=short-unkillable ($HOME_DIR is my home directory on mila cluster)

Describe the bug

When I try to run the mila code command, it exits with code 1 with the error: salloc: error: QOSMinCpuNotSatisfied salloc: error: Job submit/allocate failed: Job violates accounting/QOS policy (job submit limit, user's size and/or time limits)

Screenshots

image

Desktop

please complete the following information):

shrutij01 commented 6 months ago

okay, solved it: needed to request exactly 24 cpus with a short-unkillable partition which wasn't obvious to me from the below section in the documentation. image