June 7, 2023
Updated on February 20, 2024
You need UVa VPN to access Rivanna. The standard one “UVa Anywhere” should work. Then, login to the server using the following command
ssh [computing-ID]@rivanna.hpc.virginia.edu
with the password for your computing ID
Before running anything, you need to load some relevant modules as you did in the CS department servers. For Rivanna, the module names are slightly different
module load anaconda cuda cudnn
The first time you login, it may ask you to pick a shell. For linux systems, by default, I recommend bash
.
The default Bash configuration looks a little strange to me, so I did the following changes in my local .bashrc
to make it similar to other bash shell
PS1="\\u@rivanna:\\W\\$ "
While you are running experiments on Rivanna, keep an eye on the remaining credits left. You can find out the latest remaining credits, by using the following command
allocations
Here is an example slurm script that I used to run the some LLM evaluation