Overall Workflow

The systems on the HPCs are all Ubuntu, which is a Linux operating system. The users should have the basic knowledge about using the Linux system. Information can be found on the Linux Tutorial page.

To have an overall understanding of how to run jobs on the HPCs, a workflow is shown below:

Step	Work	Reference
1	Connect to HPC	Page
2	Create a work directory of the job under /raid/home/user_name with needed codes and dataset	Page
3	Set up the computing environment (usually by using the Docker container image, Jupyter Notebook can also be used in Docker)	Page1, Page2
4	Check the status of the HPC to decide the resources for running the job	Page
5	Submit the job (usually with specified GPU or CPU)	Page
6	Monitor the job to avoid invalid runs	Page
7	Backup your data if necessary	Page