Release every night. Run on the cluster.
We build release systems and move scientific code onto HPC clusters. Hand us the problem. We hand back a system your team runs.
Release engineering
One build for Windows and Linux. Every release carries its test report. Your developers write features; the pipeline ships them.
HPC and scientific code
Your model leaves the workstation. Its parameter sweeps run as batch jobs on your cluster, as many jobs as the sweep needs.
We keep your verified Fortran and C++ and use GPUs only where measurement says they pay.
Our cluster work runs on PBS. The same approach carries to Slurm and cloud clusters.
Proof
Asked only for “builds” on a large government simulation framework, we delivered an automated release system in about a month. It builds every night. Adding an operating system, RHEL 9, took about a day.
We also build and run our own commercial software platform in production:
What you keep
The system, plus documents and runbooks your team runs without us. You choose: we hand the system over, or we keep running it.
How to start
Consultations are free. Tell us about the problem and we’ll set up a call to work through it together.
We can start as soon as you need and join work already under way.
Who owns the work?
You do. We sign your NDA, and the code, documents, and runbooks we write for you are yours.
Do you work remotely or on site?
Both, as needed.
Do you work in our CI and our repositories?
Yes. We build the system where your team works, so it stays there when we leave.
How do we get started?
With a free consultation. We get on a call, get to know your team, and work through the problem together.