MPICH subsets run on every push and PR. OpenMPI and GPU subsets are available via manual dispatch. Each test builds in double precision, runs 3 perturbed ensemble members (4 MPI ranks), and validates with PyCECT.
Thanks to Teo Price-Broncucia and Allison Baker for their help on ensemble consistency testing. Methodology and background are given in Price-Broncucia et al. (2025), Geosci. Model Dev., 18, 2349–2372.
| Compiler | MPI | Target | Status | Container |
|---|---|---|---|---|
| GNU | MPICH | CPU |  |
hpcdev:almalinux9-gcc14-mpich-26.02 |
| GNU | OpenMPI | CPU |  |
hpcdev:almalinux9-gcc14-openmpi-26.02 |
| Intel | MPICH | CPU |  |
hpcdev:almalinux9-oneapi-mpich-26.02 |
| Intel | OpenMPI | CPU |  |
hpcdev:almalinux9-oneapi-openmpi-26.02 |
| NVHPC | MPICH | CPU |  |
hpcdev:almalinux9-nvhpc-mpich-26.02 |
| NVHPC | OpenMPI | CPU |  |
hpcdev:almalinux9-nvhpc-openmpi-26.02 |
| NVHPC | MPICH | GPU |  |
hpcdev:almalinux9-nvhpc-mpich-cuda-26.02 |
| NVHPC | OpenMPI | GPU |  |
hpcdev:almalinux9-nvhpc-openmpi-cuda-26.02 |
Compile-only workflows verify the NVHPC + OpenACC + CUDA toolchain by building on a Github Action runner without a GPU
| Compiler | MPI | Target | Status | Container |
|---|---|---|---|---|
| NVHPC | MPICH | CUDA (compile) |  |
hpcdev:almalinux9-nvhpc-mpich-cuda-26.02 |
Bit-for-bit (BFB) workflows compare history output in single precision for CPU runs (240km case; see the BFB section in .github/ci-config.env).
| Test | Status |
|---|---|
| BFB: I/O (SMIOL vs PIO) |  |
| BFB: Decomposition (1 vs 4 ranks) |  |
| BFB: I/O GPU (SMIOL vs PIO) |  |
| BFB: Decomposition GPU (1 vs 4 ranks) |  |
| BFB: NVHPC CPU vs GPU |  |
| Code coverage |   |
Container images are from ncarcisl/hpcdev.
Image tags, compiler mappings, and MPI flags are configured in .github/ci-config.env.
The Model for Prediction Across Scales (MPAS) is a collaborative project for developing atmosphere, ocean, and other earth-system simulation components for use in climate, regional climate, and weather studies. The primary development partners are the climate modeling group at Los Alamos National Laboratory (COSIM) and the National Center for Atmospheric Research. Both primary partners are responsible for the MPAS framework, operators, and tools common to the applications; LANL has primary responsibility for the ocean model, and NCAR has primary responsibility for the atmospheric model.
The MPAS framework facilitates the rapid development and prototyping of models by providing infrastructure typically required by model developers, including high-level data types, communication routines, and I/O routines. By using MPAS, developers can leverage pre-existing code and focus more on development of their model.
This README is provided as a brief introduction to the MPAS framework. It does not provide details about each specific model, nor does it provide building instructions.
For information about building and running each core, please refer to each core's user's guide, which can be found at the following web sites:
Within the MPAS repository, code is laid out as follows. Sub-directories are only described below the src directory.
MPAS-Model
├── src
│ ├── driver -- Main driver for MPAS in stand-alone mode (Shared)
│ ├── external -- External software for MPAS (Shared)
│ ├── framework -- MPAS Framework (Includes DDT Descriptions, and shared routines. Shared)
│ ├── operators -- MPAS Opeartors (Includes Operators for MPAS meshes. Shared)
│ ├── tools -- Empty directory for include files that Registry generates (Shared)
│ │ ├── registry -- Code for building Registry.xml parser (Shared)
│ │ └── input_gen -- Code for generating streams and namelist files (Shared)
│ └── core_* -- Individual model cores.
│ └── inc -- Empty directory for include files that Registry generates
├── testing_and_setup -- Tools for setting up configurations and test cases (Shared)
└── default_inputs -- Copies of default stream and namelists files (Shared)
Model cores are typically developed independently. For information about building and running a particular core, please refer to that core's user's guide.