-
Notifications
You must be signed in to change notification settings - Fork 2
Basic Knowledge
Just follow the official guide. If you are a NYU student, you do need a faculty sponsor to apply for an account.
The most frequently used liunx commands are covered in Hammond Liu's HPC Notes.
If you want to learn more, there are lots of linux beginner guide online. Pick anyone you like, for example the cheat sheet from GeeksforGeeks. Or just ask chatGPT, it knows everything.
Vim is the default text editor in HPC. It's not mendatory, but learning vim will make life much easier. I personally recommend this guide. You can also learn vim through playing a video game
One of the killer features in vscode is being able to connect to a remote server easily. Even though I'm a fan of sublime and zed, I turn to vscode when using HPC.
Open terminal app on your mac/windows/linux, and type
ssh <netid>@greene.hpc.nyu.edu
Replace <netid> with your own. You will be prompted to enter your password, which is the same as your nyu account. (Your password won't be shown on the screen, but it's still being typed in.)
If you see your terminal prompt changes to netid@log-n:~$, then you are successfully logged in.
Step1:
Install the Remote - SSH extension in vscode. You can find it by searching Remote - SSH in the extension marketplace.
Step2:
Open command palette (using F1) and type Remote-SSH: Connect to Host.... You will see a list of hosts that you have connected to before. Click on Add New SSH Host..., then type in
<netid>@greene.hpc.nyu.edu
Replace <netid> with your own. You will be prompted to enter your password, which is the same as your nyu account.
There are four main directories in HPC:
/home/$USER
/scratch/$USER
/archive/$USER
/vast/$USER
You can always check your quota usage by typing myquota.
This is where you will land when you log in. The quota is really small so you should not store your data here.
This is where you should store your data. You have a lot of space here, but be aware that the data will be deleted after 60 days if not used. Conda environment for python should be installed in singularity overlays, which is stored in scratch. You will know what it means later.
This is where you can archive your large data. It's not often used.
Another place to store your data, ideal for high I/O data that may bottleneck on the scratch file system. I don't use it personally.
Now that you understand the basic of HPC.
Move on to the Quick Start guide to get your python code running.
Or check the Manual Setup guide to set up your own conda and singularity environment.