ThermAI · Interactive Demo

Your workload. Your next Flight.

Explore CPU or GPU compute in eight guided steps.

01ComputeILLUSTRATIVE DEMO

What would you like to run?

Choose compute, then load an example to prefill the whole Flight.

Compute type

GPU Flight

AI training, batch inference and GPU-enabled simulation.

RTX 5090 or RTX PRO 6000CPU Flight

Scientific analysis, CPU simulations and parallel jobs.

Configurable vCPU and memory

Load an example workload

Parameter sweep

Compare many independent experiments.

12 tasks · 30 min averageSimulation study

Explore six engineering design cases.

6 tasks · 2 hours averageTrain a model

One longer model-training run.

1 task · 4 hours on GPU

Parameter sweep loaded

12 tasks · G1 RTX 5090 · 16 GB RAM · 0.5 h per task

thermai/pytorch:demo · /outputs/parameter-sweep/

Flight name

These prepared examples change the tasks, resources, runtime, command and data paths. Choose a different example to see a different setup.