OneFLOW Scientific Computing
OneFLOW is an open-source codebase for computational fluid dynamics (CFD) and scientific computing. It brings together a C++ runtime, mesh generation and conversion, numerical methods, case-based verification, and engineering workflow tools.
This documentation serves three audiences: people getting OneFLOW built and running, users preparing and checking cases, and developers working on the runtime or engineering tools. OneFLOW is evolving; the code, tests, and runbooks for the revision in use determine which capabilities are currently supported. Roadmap items and build options alone do not establish implementation completeness.
Project links
Getting started
Start with the project overview, then follow the installation and build instructions for your platform.
OneFLOW — project scope and documentation map.
Install — installation and environment setup.
Repository README — current build, test, and command-line instructions.
Using OneFLOW
Mesh and engineering workflows
OneFLOW Mesh Architecture — mesh information levels, snapshots, and format adapters.
Grid Topology and Validation — topology contracts, boundary normalization, and derived connectivity.
Face Stage Pipeline — staged face write path, BC export, reorder contract, identity vs orientation.
Grid case guide — mesh cases and conversion workflows.
OneFLOW Workbench 用户手册 — Chinese user guide for the optional Workbench GUI.
Case Parameter Design and GUI Editing — case parameter files and GUI editing design.
Engineering nomenclature and semantic cross-reference — source-backed cross-software CFD terminology, variable aliases, units, and semantic caveats.
Workbench module reference — detailed current behavior, paths, and service tests.
Developer resources
Test Inventory and Contract Data — source-discovered test catalog, separated by framework, with input/expected/result contracts.
OneFLOW Runtime Architecture and Development Guide — runtime entry points, configuration boundaries, execution flow, and refactoring rules.
Runtime Case Lifecycle and Production Entry — production CLI-to-case execution sequence, configuration ingestion, cleanup ownership, and integration evidence.
Database Architecture and Data Access Contracts — parameter ownership, type-erased storage, const access, field pointer registration, and serialization contracts.
Evolutionary Scientific-Computing Architecture — computation-unit contracts, intermediate representations, and the long-term algorithm/dataflow evolution model.
OneFLOW Architecture Review and Development Principles — whole-system CFD/CAE architecture checklist, major cross-cutting risks, and AI-assisted development review prompts.
OneFLOW Architecture Priority Roadmap — ranked investment sequence by system-wide return, verification cost, and cross-subsystem leverage.
Verification and Evidence Lifecycle — how claims, tests, runs, artifacts, numerical results, and references form an auditable evidence chain.
AI project context and refactoring handoff — current focus, verified findings, open gaps, and next actions for continuing work.
OneFLOW Workbench Architecture — Workbench concepts, service boundaries, run lifecycle, and shared Parameter contracts.
Verification Contract Workbench proposal — proposed shared model for numerical, geometry, runtime, workflow, and case evidence; not yet an implementation claim.
Regression test guide — CFD and grid regression checks.
Documentation organization plan — document ownership and consolidation status.
Runtime refactoring roadmap — staged goals, acceptance criteria, and implementation evidence.
Reference library
The existing reference library covers programming languages, build systems, parallel computing, numerical methods, visualization, and related tools. It remains available below; the curated entry points above are intended to help readers find a starting point without replacing that material.