LLM Cluster
Plan distributed training and inference, then open the complete same-origin simulator.
Preparing this page…
Your current navigation and account state stay available while the next page loads.
Open a lab and run the simulation in this browser. Every run is disposable, resettable, and local to the tab. No account, container, cluster, or cloud resource is created.
Each lab loads only its own runtime after you open it. Progress stays in this browser and is not saved to an account.
Plan distributed training and inference, then open the complete same-origin simulator.
Open the complete system-design simulator or trace a bounded retry-overload exercise.
LLM Cluster is based on the LLM Cluster Simulator, and Breakscale is based on the Breakscale system-design simulator. Both remain independent upstream projects.
Maintained for Frontier Rehearsal by frontier-rehearsal. No affiliation with or endorsement by the upstream authors.
These labs run deterministic or engine-backed simulations in your browser. They do not create real databases, clusters, GPUs, queues, or network services, and their results are not production benchmarks.