Computing Resources


This page lists computing resources that may support your course research project. We provide CPU VMs and facilitate access to available campus resources, including GPUs and storage, but we do not provide credits for paid AI services.

If none of the options below meets your team’s needs, consider a CPU-friendly approach, use smaller models or datasets where appropriate, or identify a fallback before relying on GPU access or paid services.

Ultimately, plan your project around resources that your team can access reliably, and be resourceful!

Acceptable Use Policy
  • Use resources obtained through the course only for course-related work.
  • Request only the resources you need, and release them as soon as you finish using them.

You must comply with the policy above. Violations may result in the immediate revocation of access.

CPU virtual machines

CloudLab provides on-demand access to computing environments, most notably CPU VMs, for research and education. We have created a CloudLab project for this course, ml4sys. Use the project invitation link to request access.

When using these resources, you must also comply with the official CloudLab Acceptable Use Policy. Refer to the CloudLab documentation for usage instructions. For technical support, contact the course staff first rather than CloudLab.

Illinois Campus Cluster

The Illinois Campus Cluster is a campus-wide high-performance computing and storage system that supports both batch jobs and interactive sessions. Job wait times are reportedly long overall, so plan accordingly and submit your jobs well in advance!

For this course, we have requested access to the shared Engineering Instructional queue/partition, which provides both CPU and GPU nodes as described below. All students on the course roster should automatically have access to the partition through the course; if not, contact the course staff for help.

Engineering Instructional partition

This partition eng-instruction consists of five compute nodes, providing a total of 512 CPU cores, 3.5 TB of RAM, and 6 NVIDIA A10 GPUs.

Node typeNodesCPUs per nodeMemory per nodeGPUs per nodeInterconnect
CPU32× AMD EPYC 7713 64-core1 TBHDR InfiniBand
GPU22× Intel Xeon 8358 32-core256 GB3× NVIDIA A10HDR InfiniBand
Access and job submission
  1. Review the Campus Cluster Getting Started guide.
  2. Confirm that your engineering account membership includes the course account by running:
    /projects/illinois/eng/shared/shared/examples/my-accounts-eng
  3. Use the provided sample batch script as a starting point:
    /projects/illinois/eng/shared/shared/examples/sample.sbatch
  4. Submit jobs to the eng-instruction partition using the account for this course 26fa-cs598fyy-eng. See the Running Jobs guide for comprehensive submission instructions.
Additional Campus Cluster documentation

Illinois Computes

Illinois Computes provide additional computing resources but they are currently intended to support research rather than instructional use. This course will not have an Illinois Computes allocation, but you may obtain access through a research group, subject to that group’s policies.

Illinois Computes Research Notebooks (ICRN)

Illinois Computes Research Notebooks (ICRN) is a no-cost, convenient option for interactive model training, providing web-based access to Jupyter Notebooks, VS Code, and PyTorch. Each notebook receives guaranteed access to one CPU core and shared access to GPU resources. If this suits your project's needs, check whether you can access ICRN.

AI-service credits

This course does not plan to provide credits for paid AI services. If your project depends on a paid service, make sure your team has reliable access and a practical fallback.