Copyright 2025 Fluid Numerics LLC
The Galapagos Studio is a fully managed Slurm cluster shared with other studio members. Fluid Numerics aims to provide AI/ML/HPC enthusiasts with access to a diverse array of hardware for testing and experimentation purposes.
We are actively going through security, billing, and legal audits. More information coming soon! We can't wait to get you on Galapagos!
On the Galapagos Cluster, you pay a flat fee for a fluidnumerics.com account, which provides you access to the cluster with 100GB /home directory space, an active Slurm account, access to the #galapagos-support slack channel. This fee is meant to cover the costs of a Google Workspace Premium license and a Tailscale enterprise license for access to our tailnet.
Beyond this, Fluid Numerics has integrated Slurm accounting with an in-house billing system that bills compute node usage per second. Each month, we compile your usage of the cluster using Slurm's built in accounting mechanisms to capture accumulated cpu-hour, gpu-hour, and mem-gb-hour consumption for each node on the cluster. Because each system on our cluster is unique and has a different power and maintenance profile, the rates for consumable resources varies with each compute node. Once costs are accumulated, our system autocharges a card on file and issues an invoice and charge receipt to you. Additionally, we provide detailed usage reports that give you visibility into the cost breakdown for each job you ran during the month.
If you are an active studio member, you can open an issue on this repository, or reach out to us on the #galapagos-support channel on Slack. Plus, we'll host regular live webinars (with Q&A) that cover basic slurm user training, usage of popular applications, and cluster updates.
We want to work with our studio members to alleviate the system administration burden so that you can simply focus on testing and optimizing your applications on a wide range of software.
If you want new/different software, just ask.
System behaving weird? Tell us.
Want to comiserate on issues related to AMD GPUs? We've got you!