GPUs, open to students
The College of Engineering AI Makerspace gives UNL engineering students access to powerful computing resources for exploring, developing, and experimenting with artificial intelligence and machine learning at multiple levels.
The AI Makerspace is part of the PRAIRIE Initiative and is a collaborative effort across the UNL College of Engineering, Scott Data Center, and Holland Computing Center. It leverages the National Research Platform to provide a variety of AI tooling and expanded compute capabilities, with a core focus of cultivating AI knowledge and talent right here in Nebraska.
Learn. Build. Expand.
The Makerspace includes 8 dedicated NVIDIA H100 GPUs, specifically for students to have access to high-performance AI computing for any number of uses:
- AI and machine learning experimentation
- Research and prototyping
- Generative AI and large language models
- Independent projects and exploration
- Class projects and coursework
Through the National Research Platform (NRP), students can also take advantage of expanded and opportunistic access to additional computing resources when available.
See the NRP Resources page for more information.
The Makerspace is available to all College of Engineering students.
Label | Value | Detail |
|---|---|---|
GPU 0-7 | NVIDIA H100 | 80 GB VRAM each |
Location | Scott Data Center | UNO Scott Campus |
Operated by | Holland Computing Center | on the NRP |
Interface | JupyterLab | browser, no setup |
Storage | 25 GB | expandable on request |
Cost to you | $0 |
|
Getting Access
1: Complete the training
A self-paced course on Bridge covering the hardware, how to launch a session, and what you're responsible for on a shared system. Enroll yourself — no approval needed to begin.
Open the course on Bridge2: Read the documentation
HCC's Makerspace guide covers launching a session, moving data with Globus, and the models available to you. Worth reading before your first login.
Read the text3: Sign in and launch
Your university email is added to the access list after you finish the course. Then you sign in, pick your resources, and your session starts in the browser.
Launch your sessionWhat you can do with it
More than one kind of resource.
Not every project needs a GPU, and the fastest path is often the one that doesn't take one.
Whole H100 GPUs
Train and fine-tune – 80 GB of VRAM per GPU lets you work with models and datasets that won't fit on a laptop or a consumer graphics card. Best suited to training runs, fine-tuning, and computer vision work.
Hosted models · no GPU
Use large language models. The National Research Platform hosts frontier-scale models — including large general, multimodal, and embedding models — free to university users through a web chat interface. If a hosted model does the job, you don't need to allocate anything.
CPU sessions
Prepare data and write code. Downloads, cleaning, preprocessing, and debugging don't need a GPU. CPU-only sessions start faster and leave the GPUs free for work that actually uses them.
What's expected of you
Request what you need.
A GPU you're holding but not using is one another student can't get. Start smaller than you think you need — you can always stop and relaunch with more.
Stop your session when you're done.
Closing the browser tab doesn't stop the server. Idle sessions are reclaimed automatically, but that's a backstop, not a plan.
Clean up your data.
Old datasets and checkpoints stay until you delete them. Managing your own space is easier than having someone else decide what's expendable.
Your account is yours.
Everything run under your credentials is your responsibility. University computing policy applies here, and so does the National Research Platform's.