General concepts
How meshes, finite element fields, weak forms, and solvers fit together.
A Rodin application connects a geometric domain, a discrete field space, a weak problem and an algebraic solver. These guides develop those objects in that order and identify the C++ interfaces used at each stage. The examples and tutorials supply complete numerical problems; the module reference describes individual APIs.
Choose a reading route
| Task | Starting point | Continue with |
|---|---|---|
| Build and run a first solve | Installation and setup | A first problem |
| Understand the finite element construction | Ciarlet's definition of a finite element | Finite element spaces and Weak formulations |
| Translate a model into C++ forms | Variational formulations in Rodin | Boundary conditions and Solvers |
| Inspect a computed field | Input, output, and visualization | Gallery and Examples and tutorials |
| Move a local problem to a distributed backend | Backends and supported configurations | Distributed computing with MPI and Parallel solvers and distributed linear algebra with PETSc |
Start with a complete problem
The introductory sequence assumes basic C++ and familiarity with calculus and linear algebra. Detailed finite element background is introduced in the guides below.
- Installation and setup — Configure the dependencies, build Rodin, and verify the installation.
- First steps with Rodin — Build a small program and identify the roles of a mesh, a finite element space, and symbolic fields.
- A first problem — Build and run a Poisson solve, then inspect the exported solution.
- Core concepts — Follow the complete path from geometric data to the assembled algebraic system.
Describe the domain
A mesh records both geometric coordinates and topological relations. These guides establish the distinction before introducing integration and fields. Incidence relations describe adjacency; geometric maps supply coordinates, Jacobians and measures for evaluation and integration.
- Meshes — Create, load, inspect, and traverse a mesh.
- Polytopes — Cells, faces, reference elements, and mappings.
- Connectivity — Incidence relations between vertices, edges, faces, and cells.
Approximate fields on the mesh
A finite element specifies local approximation and degrees of freedom; identifying compatible degrees of freedom across cells produces a global space. Field range, geometric mapping and coefficient ordering are separate choices in this construction.
- Ciarlet's definition of a finite element — The finite element triple and local degrees of freedom.
- Finite element spaces — Scalar, vector, and complex spaces and grid functions.
- Quadrature formulas — Numerical integration, exactness, and curved cells.
- Matrix storage and finite-element component ordering — Component numbering, tensor indexing, and storage.
Express and solve equations
The continuous weak problem determines the forms, boundary conditions, and algebraic properties that a solver must respect. Follow a derivation before selecting the corresponding form expressions, then check constraints, quadrature and solver status against that problem.
- Weak formulations — Derive weak forms from Poisson and elasticity equations.
- Variational formulations in Rodin — Express bilinear and linear forms in C++.
- Boundary conditions — Essential, natural, periodic, and mixed boundary conditions.
- Solvers — Choose and configure linear and nonlinear solvers.
- Numerical contracts — Signs, projection versus interpolation, constraints, quadrature, and solver lifetimes.
- Interfaces, traces, and discontinuous Galerkin — Traces, facet integrals, jumps, and averages.
Inspect and exchange results
Mesh topology, geometric coordinates and field coefficients must be exported together with their interpretation. The I/O guides distinguish stored data, visualization output and time-dependent snapshots.
- Input, output, and visualization — Select mesh and field formats, export to ParaView, and write time series.
Build application models
These guides combine the preceding concepts into fitting, design evolution and constitutive models. Their application examples expose geometry checks, objective histories and residual or derivative checks appropriate to each workflow.
- Fitting a mesh interface with SWIFT — Fit meshes to marked interfaces and inspect quality.
- Shape and topology optimization — Shape derivatives, regularization, and boundary evolution.
- Solid mechanics — Finite-strain kinematics, constitutive laws, residuals, and consistent tangents.
Choose an execution backend
The local workflow provides the starting point. Parallel execution adds ownership, communication, and synchronization requirements. Establish the discrete problem locally before following the backend-specific construction and collective-operation rules.
- Backends and supported configurations — Local, OpenMP, MPI, and PETSc execution paths.
- Distributed computing with MPI — Distributed meshes, partitioning, and reconciliation.
- Parallel solvers and distributed linear algebra with PETSc — Distributed algebra, KSP/SNES, and multi-field problems.
- Thread safety and parallelism — Shared state and synchronization requirements.
Further reference
The notation reference defines mathematical symbols. The module reference provides API declarations, and the project repository hosts source code and issue discussions.