I built a physics computation platform that replaces brute-force numerical PDE solvers with analytic geometry evaluated in a 13D coordinate space. Instead of discretizing onto million-cell meshes and iterating to convergence on a supercomputer, GDBS maps physical systems into geometric positions where observables become exact coordinates. Evaluation is O(N) instead of O(N³), so what takes HPC clusters hours runs in <100ms in WebAssembly in the browser.

7 physics domains: plasma/fusion, materials science, cosmology, geophysics, fluid dynamics, quantum information, and molecular/medical. 300+ validation tests against NIST, CODATA, Planck 2018, ITER Physics Basis, PREM, and peer-reviewed literature. Example: bulk modulus of diamond — GDBS: 443 GPa, CRC Handbook: 442 GPa.

The stack is Rust→WASM for the physics engine (6,300+ LOC, 59 modules), .NET 8/C# backend, vanilla JS frontend. Includes a query language (GQL), REST API, and a built-in precision system that tracks numerical drift across 65 orders of magnitude (validated at 0.27% error on Hawking radiation calculations spanning ℏ at 10⁻³⁴ to thermodynamic scales). Free 14-day trial with full access. Theory module is free permanently. Beta program available for researchers (citation required).

- https://gdbs.getvaultsync.com - https://doi.org/10.5281/zenodo.18753425 (Geometric Number Scale - PrePrint)

I do hope it helps.