Basic Training: Beginner Crash Courses¶
Build the skills at the heart of productive work on Deceema: navigating its Linux environment, automating safely with Bash, writing a reproducible analysis, and asking Slurm for the right compute resources.
No prior HPC experience is required. Every track starts with the vocabulary, shows one concept at a time, and gives you a small working result before moving to Deceema-scale computation.
How to use these tutorials
Type the examples yourself in a disposable training directory. Read each command before pressing Enter, compare the result with the explanation, and stop when the output differs from what you expect. The goal is understanding—not racing to the final command.
-
Linux (RHEL)
Navigate the filesystem, inspect data, understand permissions, manage software environments, and work safely on a shared Linux platform.
-
Bash
Turn commands into reliable scripts with careful quoting, arguments, pipelines, validation, and failure handling.
-
Python
Create an isolated environment, write a command-line program, and run it as a reproducible Deceema job.
-
R
Turn an interactive analysis into a script with explicit inputs, outputs, and a captured software environment.
-
Slurm
Understand resource requests, submit a batch script, follow it through the queue, and inspect its result.
Recommended Learning Path¶
- Begin with Linux (RHEL) to learn the environment and safe file operations.
- Continue to Bash to automate repeatable command-line work.
- Choose Python or R and complete its analysis exercise.
- Complete Slurm to move that script onto Deceema compute resources.
- Use Jobs on Deceema for arrays, dependencies, GPU jobs, accounting, and deeper troubleshooting.
Before You Begin¶
You should have:
- An active Deceema account and access to a login node.
- A project code and corresponding Quality of Service (QoS), confirmed from your Deceema access approval or account details.
- A working directory for scripts, input data, logs, and results.
- Basic familiarity with navigating directories and editing a text file from the command line.
Tip
Training works best with small, disposable input data. Prove the workflow at small scale before applying it to important or large project data.
What Good Looks Like¶
By the end of Basic Training, you should be able to:
- Run a Python or R script from the command line without manual interaction.
- Navigate RHEL confidently and inspect files without changing them accidentally.
- Write a Bash script that validates inputs and stops when a command fails.
- Separate source code, configuration, input, logs, and results.
- Record the language version and packages used by a run.
- Write a Slurm script with matching account and QoS values.
- Submit, monitor, cancel, and inspect a job by its ID.
- Recognize when to consult Storage, Service Status, or Support.