VanellusGPU-accelerated thermal simulation. Sign up

Features

Thermal simulation, accelerated.

Model airflow and heat transfer in electronic hardware, run simulations on cloud GPUs, and retrieve results programmatically.

GPU performance

Order-of-magnitude speedups with GPU compute.

Run GPU-accelerated CFD and conjugate heat transfer for electronics cooling, with faster feedback on airflow and component temperatures.

See the validation →
  • Performance

    GPU-native CFD solver

    Finite-volume methods and linear solvers designed for parallel GPU execution, accelerating every simulation.
  • Physics

    Thoroughly validated

    Coupled airflow and solid conduction, checked against experiments and established CFD tools.
  • Compute

    Cloud HPC

    Simulations run on high-performance cloud GPUs, with no on-premises hardware to buy or maintain.

Programmable cloud simulation

Run one case or a complete design study through the same API.

Submit JSON from Python or any HTTP client. Each simulation runs asynchronously on a cloud GPU, so you can launch variants in parallel without managing your own HPC. 

import requestsfor design in designs:    requests.post(        "https://api.vanellus.tech/simulations",        headers={"X-API-Key": api_key},        json=design,    )
  • ToolsFits your existing workflowSubmit the same JSON request from Python, cURL, CI jobs, or optimisation scripts.
  • ComputeScales across design variantsIndependent cases run on cloud GPUs instead of competing for one workstation.
  • ResultsReturns data to your codeGet live updates on progress and key values such as temperature, then retrieve CSV and VTU results through the API.

Simulation capabilities

Comprehensive thermal simulation, built on proven CFD.

Finite-volume CFD

Model natural and fan-driven airflow around hardware.

  • Cell-centred finite-volume solver
  • Steady laminar and k–ω SST RANS flow
  • Exterior and linked fans using tabulated fan curves
Docs: What the Vanellus solver does →

Conjugate heat transfer

Solve airflow and solid conduction together to predict component temperatures.

  • Coupled solid and fluid domains
  • Volumetric and component heat loads
  • Forced and buoyancy-driven natural convection
Docs: Flow and thermal models →

Compact thermal models

Predict electronic package junction temperatures without resolving their internal construction.

  • Two-resistor and star-resistor networks
  • Thermal-resistance paths to package faces
  • Junction-temperature monitoring
Docs: Compact thermal models →

CAD and meshing

Build a simulation domain from simple primitives and production CAD.

  • STEP and STL import with selectable units
  • Cuboid primitives for rapid setup
  • Automatic meshing with per-component sizing
Docs: Geometry and mesh →

Cloud API

Submit, monitor, and automate GPU simulations from any HTTP client.

  • JSON requests from Python, cURL, or other clients
  • Asynchronous cloud GPU runs
  • Live progress, residuals, and monitor values
Docs: Run simulations →

Engineering outputs

Inspect complete fields or track the values that determine whether a design works.

  • Temperature, pressure, velocity, and turbulence fields
  • Separate fluid and solid VTU results
  • CSV histories and monitor values
Docs: Results and simulation lifecycle →

Try the API

Build simulation into your workflow.

Start simulating now with 100 free GPU-minutes.

Sign up Read the docs