Is this course for you?

Read this before signing up

This 1-credit practicum is deliberately self-selecting. It’s not a “ML for everyone” course — it’s a course about a specific skill (running research code on a shared HPC cluster) that some grad students need and others genuinely don’t. This page tries to honestly help you figure out whether you’re in the “need it” group.

If you’re already familiar with HPC, Linux, and Slurm, you can skip straight to the syllabus — this page is for people who landed here and aren’t sure what any of the terminology means.

For the deep version of “what is HPC and why does it matter?”, read the handbook’s Start here: Why HPC? page. The short version is below.

1. In one paragraph: what this course teaches

An HPC cluster is a room full of computers in a data center somewhere that someone else manages for you. You log in remotely from your laptop and use them — for running deep-learning training, processing huge datasets, doing parameter sweeps, anything where your laptop isn’t enough. OSU has one such cluster, Unity (operated by the College of Arts and Sciences). It’s free for OSU students. This course teaches you how to use it productively — from your first SSH connection to a complete reproducible research repo running real code on the cluster — in 15 weekly labs, each taking about an hour.

2. Is HPC actually something you need?

✔ You probably want this course if:

  • Your laptop runs your code for hours or days at a time, blocking you from other use
  • You work with deep neural networks (CNNs, transformers, LLMs) and your laptop lacks an NVIDIA GPU — most non-gaming laptops do
  • You work with satellite imagery, climate model output, large genomics data, video, or other inherently large datasets
  • You want to run many similar experiments in parallel (hyperparameter sweeps, ensemble simulations, etc.)
  • Your lab has shared storage on Unity (BuckAI has 1 PB) and your collaborators expect you to use it
  • Your advisor has mentioned “the cluster” or “Unity” or “Slurm” more than once

❌ You probably don’t need this course right now if:

  • Your work is mostly writing, theorem-proving, or literature review with light computation
  • Your code uses scikit-learn on tabular data that trains in minutes on your laptop
  • Your laptop runs everything you need to do without ever feeling slow
  • You do pure statistical analysis on a few-GB-or-less dataset
  • You’re early in your PhD and don’t yet know what kind of computation you’ll need — bookmark this and come back later

There is no shame in the second list. Many excellent PhDs are completed entirely on a laptop.

3. “But I’ve only ever used Windows”

That’s fine. HPC clusters run Linux, which is neither Windows nor macOS, but Windows users get there via a small one-time setup (WSL2) that gives you a Linux terminal on your Windows machine. After that one-time setup, the rest of this course applies to you the same as anyone else.

Your day-to-day editing happens in VS Code (which runs natively on Windows, macOS, and Linux), so you’re never living inside Linux — Linux is just the OS on the remote cluster you connect to.

You do not need to:

  • ❌ Install Linux on your laptop
  • ❌ Switch away from Windows
  • ❌ Give up familiar tools

You do need to:

  • ✔ Spend ~10 minutes setting up WSL2 (one time)
  • ✔ Get comfortable with a small set of Linux commands (cd, ls, cat, editing a file) — most of this you probably already know

4. What you’ll spend time on

The course is 1 credit, structured as ~1 hour/week for 15 weeks (with occasional 1.5-hour weeks if HPC bites you, which it sometimes will). The labs build cumulatively — by week 12 you’ll have:

  • A working SSH + VS Code Remote-SSH setup with multiplexing
  • A productive .bashrc and the livenode workflow for persistent sessions
  • A working mamba environment, shared with labmates if you have any
  • A Jupyter notebook on the cluster, viewable in your laptop’s browser
  • The ability to submit, monitor, and right-size Slurm batch jobs
  • A small capstone project packaged as a reproducible repo

Each lab produces a verifiable artifact (a config file, a Slurm log, a seff efficiency report, etc.) that you can self-check against the lab’s deliverable list.

See the full syllabus for the week-by-week breakdown.

5. What this course is not

So you know what you’re not signing up for:

  • ❌ Not a course on machine learning, deep learning, or statistics. We assume you already know those (or will learn elsewhere); this course teaches the infrastructure to run them at scale.
  • ❌ Not a Linux administration course. You’ll learn just enough Linux to use the cluster — cd, ls, nano, bash aliases. We don’t cover kernel compilation, networking config, or system administration.
  • ❌ Not a Python tutorial. We assume comfort with Python (functions, packages, file I/O); we teach how to run Python at scale via Slurm.
  • ❌ Not an Anthropic Claude course. AI coding assistants are part of the workflow (you’ll pick one of Claude Code / GitHub Copilot / Gemini Code Assist in Lab 2), but they’re tools, not the subject.

6. What you’ll need

  • Laptop: macOS, Linux, or Windows (the last with WSL2). The instructions are macOS-flavored but every step has Windows/Linux equivalents.
  • OSU account: you need an @osu.edu email and a Unity account. Account requests are free and typically processed within 1–3 business days — request through your department, your advisor, or asctech.osu.edu. (You can also use OSC accounts as alternative — same techniques apply.)
  • Time: ~1 hour/week of focused, undistracted work. HPC has more “wait for it to schedule” time than typical coursework, so part of the budget is patience.
  • An AI coding assistant (free options exist): Claude Code via claude.ai, GitHub Copilot via GitHub Education (free for verified students), or Gemini Code Assist through your OSU Google account.

7. Decision points

After reading this page, your honest options are:

  1. “Yes, this looks like it’ll save me time later.” → Continue to the syllabus and start Lab 1.
  2. “I’m not sure but curious.” → Read the handbook’s Why HPC? page for the longer answer, and skim Lab 1 to see what week 1 actually looks like.
  3. “Doesn’t sound like me right now.” → Bookmark this. Come back when your research direction changes or your laptop starts feeling slow.

All three are valid responses. The course is designed for option 1 students, but option 2 readers often discover they’re actually option 1 once they see what HPC concretely unlocks.

What’s next