On the University of Copenhagen, the Faculty of Science is hosting infrastructure for use in research and education. The entry to these services is called Electronic Research Data Archive, ERDA. As a student, you can log in to the service with your normal UCPH credentials, and use the storage space for collaborations or datasets.
When you are working on datasets stored on ERDA you can use the DAG service, which is implemented as a Jupyter notebook where the ERDA storage is available. The Python interactive environment can be a great way to explore large datasets without having to download the dataset to your machine.
For the HPPS course, you can use DAG and start a terminal to get access to a working GCC compiler as well as Make and other tools. When you use DAG, you get a shared server, suitable for exploration and testing, but not suited for long running sessions, or performance measurements.
For larger compute tasks, you can use the more performance oriented MODI service. This service has a set of reasonably powerful nodes that can be used to test and benchmark your solutions. It is intended to be a testing site for HPC applications, and is thus ideally suited for the 4th and final assignment in the course.
Like most shared, large-scale facilities, MODI only allows execution of jobs submitted through the queue system. Similar to large-scale facilities, MODI uses the SLURM scheduling system to give multiple users exclusive fair-usage allocations.
While SLURM contains many features, the most common ones are covered in a SLURM Quickguide.
Important: When running on MODI all your files need to be placed in ~/modi_mount as this is the only folder the execution nodes can see. You need to place both the executable and any input/output files here. You can also only submit jobs from this folder, as the results cannot be written elsewhere.