On the recordRead 9 Oct 202620 facts
PetIBM
Plans and prices
PetIBM
Free
Open-source solver toolkit; Linux officially supported; requires PETSc and MPI
Read 9 Oct 2026 · barbagroup.github.io
On the record
- API
- PetIBM installs a library and header files for building flow solvers, with API documentation and simplified example solvers provided.barbagroup.github.io · read 9 Oct 2026
- Assessment tools
- Nobarbagroup.github.io · read 24 Sep 2026
- Audience
- The project provides end-user installation instructions and separate guidance for developers extending or debugging PetIBM.barbagroup.github.io · read 9 Oct 2026
- Authoring tools
- Nobarbagroup.github.io · read 24 Sep 2026
- Bodies and meshes
- It supports multiple immersed bodies, prescribed motion, and 2D and 3D stretched Cartesian meshes.barbagroup.github.io · read 9 Oct 2026
- Compatibility limit
- The CMake instructions for finding PetIBM dependencies exclude MPI, which must be found separately.barbagroup.github.io · read 9 Oct 2026
- CPU and GPU solvers
- Linear systems can be solved on CPUs with PETSc KSP objects or on multiple CUDA-capable GPUs using NVIDIA AmgX, connected through AmgXWrapper.barbagroup.github.io · read 9 Oct 2026
- Delivery mode
- desktopbarbagroup.github.io · read 24 Sep 2026
- Dependencies
- Building from source requires a GNU C++ compiler, PETSc with parallel HDF5, an MPI implementation, yaml-cpp, and SymEngine; AmgX and AmgXWrapper are optional for GPU linear solvers.barbagroup.github.io · read 9 Oct 2026
- Installation
- End users can install prebuilt GPU-enabled packages with Anaconda or use Docker and Singularity/Apptainer containers.barbagroup.github.io · read 9 Oct 2026
- I/O
- The feature list includes HDF5 input/output support.barbagroup.github.io · read 9 Oct 2026
- Lesson library
- Nobarbagroup.github.io · read 24 Sep 2026
- License
- PetIBM code is under the BSD 3-Clause License.barbagroup.github.io · read 9 Oct 2026
- Methods
- It includes the Immersed Boundary Projection Method and a decoupled version of that method.barbagroup.github.io · read 9 Oct 2026
- Offline access
- Yesbarbagroup.github.io · read 24 Sep 2026
- Parallel computing
- It supports distributed-memory architectures, multiple GPUs on one node, and GPU clusters.barbagroup.github.io · read 9 Oct 2026
- Platform limits
- Linux is the only officially supported system; the documentation says macOS and Windows may or may not work and offers no assistance for issues on those systems.barbagroup.github.io · read 9 Oct 2026
- Purpose
- PetIBM implements immersed-boundary methods to solve 2D and 3D incompressible Navier–Stokes equations on stretched Cartesian grids using a projection approach.barbagroup.github.io · read 9 Oct 2026
