Rio
Rio (pronounced RYE-oh) is comprised of 34 compute nodes in the standard partition, providing a total of 1,560 x86 64-bit compute cores. Each HPC node is equipped with 375 GB of RAM to accommodate memory-intensive applications. Rio also includes an NVIDIA HGX H200 GPU, and additional L40s GPU nodes within the gpu partition, designed to support AI and machine learning workloads.
Open OnDemand is a web-based portal that provides access to HPC resources, including applications such as RStudio, JupyterLab, and a desktop environment for running Graphical User Interface (GUI)-based applications on compute nodes. The OOD Dashboard can be accessed by connecting to the HSVPN and opening a web browser to the IP of the VM.
Ivy
Ivy virtual machines (VMs) provide virtual computing environments specifically designed for interactive and small-scale analysis of highly sensitive data. Linux VMs on Ivy are provisioned to provide integrated access to the Rio HPC system and Open OnDemand (OOD) interface by default. Once your Linux VM is provisioned, you can seamlessly use it as a login node to Rio and access OOD without submitting a separate request.
High Security Zone Storage
Storage within the HSZ is designed for storing highly sensitive or regulated research data and to interface with the Ivy virtual environment and Rio HPC cluster.
Already have access?
System Details
Windows VMs
Virtual machines are available on two platforms: Rocky Linux 8, which by default provides access to Rio HPC resources and comes with Slurm and Open OnDemand pre-installed, and Windows Server 2019, offered as a standalone option. Each platform is available in multiple instance types. Once created, your instance will be assigned a private IP address that you will use to connect to it (in the format 10.xx.xx.xx). VMs exist in a private, secure network and cannot reach outside resources on the Internet.
Learn more about pricing and VM specs
Rio Hardware Configuration
Currently, Rio comprises of 34 compute nodes in the standard partition, providing a total of 1,560 x86 64-bit compute cores. Each HPC node is equipped with 375 GB of RAM to accommodate memory-intensive applications. Rio also includes an NVIDIA HGX H200 GPU, and additional L40s GPU nodes within the gpu partition, designed to support AI and machine learning workloads.
| System | Cores/Node | Memory/Node | Specialty Hardware | GPU Memory | GPU Devices | # of Nodes |
| Rio | 96 | 2TB | GPU: H200 | 141GB | 8 | 1 |
| Rio | 40 | 375GB | - | - | - | 34 |
| Rio | 32 | 750GB | GPU: L40s | 48GB | 8 | 4 |
Rio | 72 | 480GB | GPU: GH200 | 144GB | 1 | 32 |
Job Partitions
Rio is a managed resource. Users must submit jobs to queues controlled by a resource manager, also known as a queueing system. The manager in use on Rio is Slurm. Slurm refers to queues as partitions because they divide the machine into sets of resources. There is no default partition, and each job must request a specific partition. Partitions and access policies are subject to change, but the following table shows the current structure.
| Partition | Purpose | Max time / job | Max nodes / job | Max cores / user | Max cores / node | Default memory / core | Max memory / node |
| standard | CPU-based jobs | 7 days | 10 | 952 CPU | 40 | 4 GB | 375 GB |
| gpu | GPU jobs | 3 days | 1 | 22 GPU | 96 | 4 GB | 750 GB |
| gpu-h200 | larger GPU jobs | 3 days | 1 | 8 GPU | 96 | 4 GB | 2 TB |
| gpu-gh200 | Specific Arm architecture | 3 days | 16 | 8 | 72 | 4 GB | 480 GB |
Scheduled Maintenance
Beginning June 17, 2026, all Ivy VMs will undergo scheduled maintenance every two weeks on Wednesdays. During these maintenance windows, VMs may be restarted or temporarily unavailable.
- Overnight Maintenance Window: Wednesday, 2:00-3:00 a.m.
- Daytime Maintenance Window: Wednesday, 8:00 a.m. –12:00 p.m.
During maintenance windows, we encourage you to save your work regularly. Although unexpected, there is a chance you could experience brief interruptions or VM reboots.