CFD & HPC consultancy · India

Simulation thatstands up to review.

You send the problem. We check it against published work, choose the method, run it on high-performance computers and hand back results, files and data you can cite.

Send a problem statementHow it works

Live solver · lattice Boltzmann, D2Q9 Re 120St …Step 0

One problem statement.
Four steps.

  1. 01

    State

    You describe the problem: the geometry, the conditions and the number you need to know.

  2. 02

    Review

    We check it against published studies and benchmark cases, then propose the methods and models that fit.

  3. 03

    Compute

    The agreed analysis runs on HPC for a fixed budget of core-hours. Mesh, time step and convergence are checked, not assumed.

  4. 04

    Deliver

    Results, case files and data go to a public Zenodo record with a DOI, or to a private GitHub repository shared with your team.

Results you can cite,
or keep to yourself.

Open

Zenodo, with a DOI

A public, versioned record that can be cited in a paper, a thesis or a report, under the licence you choose.

Private

GitHub, by invitation

A private repository shared with your collaborators. Every file, log and script, under your control.

your-study/
├─ README.md
├─ CITATION.cff
├─ geometry/
├─ mesh/          quality report included
├─ case/          every solver setting
├─ logs/          residuals, monitors
├─ results/       fields, tables, figures
└─ report.pdf
Everything needed to run the study again.

Compute, measured
in core-hours.

Each project is scoped as a number of hours on parallel hardware and agreed before anything runs. You know what will be computed, for how long, and why.

  • ParallelMPI solvers across many cores and nodes
  • VerifiedMesh and time-step independence, with the evidence
  • LoggedMonitors, residuals and checkpoints kept for every run
  • OpenOpen-source solvers by default, so anyone can rerun the case

Led by Dr. Anantha Raman Lakshmipathi (opens in a new tab).

Associate Professor of Mechanical Engineering, working on conjugate heat transfer, battery thermal management and computational quantum chemistry. The studies are published with their data and code, openly, on Zenodo.

Asked often.

What do I need to send?

A problem statement is enough to start: what you are designing or studying, the operating conditions and the result you need. Geometry, drawings and data help, and can follow later.

How is a project priced?

Per scope. After the review we propose a fixed scope: the cases, the methods and the core-hours they need, and quote for that.

Who owns the results?

You do. A public Zenodo record carries the licence you choose; a private repository stays private. We sign an NDA before you share details if you need one.

Which software do you use?

Open-source solvers by default, such as OpenFOAM for flow, CalculiX and Code_Aster for structures, and Quantum ESPRESSO and PySCF for chemistry, so the case can be rerun without a licence. If your team works in a particular package, tell us.

My simulation will not converge. Can you help?

Yes. Send the case, the software version and what goes wrong. We find the cause, fix it and explain what changed.

Do you work with students and researchers?

Yes, as research assistance: help with the study design, the setup, verification and publishing the data. The research and the writing remain the researcher’s own.

Send the problemstatement.

We reply with a review of related work, an approach and the compute it needs.

Start a project[email protected]