dicklesworthstone/frankensim — explained in plain English
Analysis updated 2026-05-18
Build custom geometry, meshing, or physics solver components on a certified numerical substrate.
Track evidence and reproducibility for a simulation or optimization pipeline via the built-in ledger.
Run gradient-based or derivative-free design optimization with dimensional-unit checking.
Use individual crates like sparse matrix assembly or Chebyshev collocation as standalone libraries.
| dicklesworthstone/frankensim | sergiobenitez/ubyte | adysec/clawbot | |
|---|---|---|---|
| Stars | 36 | 36 | 37 |
| Language | Rust | Rust | Rust |
| Last pushed | — | 2023-10-02 | — |
| Maintenance | — | Dormant | — |
| Setup difficulty | hard | easy | moderate |
| Complexity | 5/5 | 2/5 | 3/5 |
| Audience | researcher | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires a pinned Rust nightly toolchain and specific sibling projects checked out at exact commit pins via a custom bootstrap tool before it builds.
FrankenSim is a large, still-in-progress Rust workspace aimed at building simulation and design optimization infrastructure where every part of the system, from physical units to numerical error to geometry validity, is tracked explicitly rather than left implicit. The stated problem is that simulation tools often let an optimization run produce a final answer without a clear, durable record of which assumptions, approximations, and machine conditions made that answer trustworthy. FrankenSim's answer is to bake tracking of units, deterministic numerical contracts, geometry evidence, and a full history ledger directly into the architecture of the codebase. The project is organized as 126 separate crates in the main Cargo workspace, plus a standalone nested workspace for WebAssembly support, covering areas such as runtime execution, numerical kernels, geometry representations like signed distance fields and meshes, physics solvers, gradient and adjoint computation for optimization, image handling, and a ledger system built on a custom SQLite-based storage layer that records artifacts, operations, and events over time. Every one of the 127 crate directories has an accompanying contract file, and the workspace includes 276 test files that exercise these contracts. There is not yet a packaged end-user simulation application or a published crates.io release. Today the project is usable as a source workspace and a library foundation for people who want to build on individual pieces, such as assembling sparse matrices, running collocation based stability probes, attaching evidence to numerical values, working with physical quantities that carry dimensional checks, or running gradient-based and derivative-free optimizers. Building it from source requires a pinned nightly Rust toolchain and specific sibling projects checked out alongside it, set up through a bootstrap tool included in the repository. This project is aimed at researchers and engineers working on scientific computing, physics simulation, or design optimization who want a rigorously tracked, evidence-based foundation to build custom solvers and geometry tools on top of, rather than a ready-to-use simulation application. The full README is longer than what was shown.
A large in-progress Rust workspace of 126+ crates for building certified, evidence-tracked physics simulation and design optimization tools.
Mainly Rust. The stack also includes Rust, Cargo, WebAssembly.
Setup difficulty is rated hard, with roughly 1day+ to a first successful run.
Mainly researcher.
This repo across BitVibe Labs
Verify against the repo before relying on details.