[General boards] [Fall 2018 courses] [Summer 2018 courses] [Winter 2018 courses] [Older or newer terms]

More machines available


#1

You can ssh remotely to coral01 - coral05.teach.cs.toronto.edu to use GPUs with your assignment. These are not bookable - first come, first served - so you’ll need to check whether others are already logged in before working on your assignment.

Regards,
Gerald Penn


#2

Things to be watch out for:

  • If someone else is on the same machine and has started using the GPU with TF, you’ll probably run into an OOM error with TF. This is because TF allocates all the GPU RAM for itself when starting a session.
  • If you find issues on some lab machines, try running your code on these machines. Some have experienced issues in one of the labs that went away on these machines.