Speedup on representative CHT problem versus OpenFOAM running on high end CPU node.
GPU-native thermal simulation for electronics cooling
Accelerate thermal design of high density electronics.
Vanellus helps thermal teams replace slow, experiment-heavy iteration with rapid, simulation-driven design. Run high-fidelity finite-volume CFD on cloud GPUs, mesh quickly, and stay in a browser-accessible workflow from setup to optimization.
Run large thermal campaigns without buying or competing for on-prem HPC.
Trusted finite-volume CFD and thermal physics, benchmarked against NASA turbulence cases.
Product
The thermal design loop, rebuilt around modern hardware
Vanellus combines automated meshing, a cloud GPU solver, and a browser-accessible interface so teams can move from geometry to decision faster.
Setup
Import PCB and CAD geometry, generate meshes automatically, and reduce setup overhead without relying on a specialist meshing workflow.
- Unified PCB and CAD workflow
- Live remeshing when geometry changes
- Wide range of thermal and fluid boundary conditions
Simulate
Run high-fidelity finite-volume CFD and thermal cases on high-end cloud GPUs with a solver architecture designed for massively parallel numerical work.
- Cloud GPU delivery by design
- Higher fidelity simulations for greater accuracy
- Faster runtime to close design-feedback loop
Optimize
Explore more design candidates before you commit to hardware, reduce costly experiments, and improve cooling performance with more confidence.
- More design exploration per project
- Lower prototype and cooling costs
- Quicker time to market
Representative benchmark
Not just GPU support. A solver stack built around GPUs from the start.
The Vanellus solver is built from the ground up to utilize high power GPUs, making design choices across meshing, memory layout and numerical solving algorithms which maximize compute throughput. This allows us to dramatically improve on the performance of CPU and CPU/GPU hybrid physics solvers.
Applications
Built for the electronics cooling problems where slow iteration kills margins.
Vanellus is aimed at thermal engineers, product designers, and team leads working on dense electronics, tighter form factors, and increasingly demanding cooling targets.
Server blades
Rigorous modelling of:
- Combined air and liquid cooling systems
- Heat pipes
- IC thermal behaviour
- Dense PCB layouts
Faster decisions on heat sinks, fans, enclosure layouts and liquid cooling systems.
Consumer electronics
Optimise:
- Consumer comfort
- Internal thermal transport with vapour chambers
- Battery performance
Reduce overspecified cooling systems and experience fewer late-stage thermal design surprises.
Data centers
Improve design and operational efficiency with:
- Reduced facility PUE
- Hall layouts
- Increased cooling fluid operating temperature
More confidence in facility and system-level cooling tradeoffs.
About Vanellus
Scientific computing depth, product pragmatism.
Why it exists
Legacy thermal tools force teams to accept runtime bottlenecks, setup complexity, local compute limits, and too much experiment-driven iteration.
Laurence Cullen
CEO · Application, meshing, and cloud infrastructure
Generalist software engineer building the product surface, geometry workflows, and cloud delivery that make Vanellus practical for real engineering teams.
LinkedIn · laurence@vanellus.tech
Dr Michael Negus
CTO · Core CFD solver
Builder of the Vanellus finite-volume CFD solver, extending the supported physics while keeping the numerical foundations rigorous and production-minded.
LinkedIn
Dr Oscar Holroyd
CSO · GPU optimization
Focused on GPU performance engineering, solver optimization, and the low-level architecture required to turn numerical methods into real throughput gains.
LinkedInNext step
Give your thermal team more design shots on goal.
Use Vanellus to run more cases, shorten iteration cycles, and ship better-performing hardware with more confidence.