Project home: https://confluence.ecmwf.int/display/ATLAS
Contact: Willem Deconinck ([email protected])
Publication:
Deconinck et al, 2017 --- Atlas: A library for numerical weather prediction and climate modelling
Atlas is a ECMWF library for parallel data-structures supporting unstructured grids and function spaces, with the aim to investigate alternative more scalable dynamical core options for Earth System models, and to support modern interpolation and product generation software
Atlas is predominantly C++ code, with main features available to Fortran codes through a F2003 interface. It requires some flavour of Unix (such as Linux). It is known to run on a number of systems, some of which are directly supported by ECMWF.
Tested compilers include:
- GCC 4.9.1, 5.3.0, 6.3.0, 7.2.0
- Intel 15.0.2, 16.0.3, 17, 18
- CCE 8.4.5, 8.5.8, 8.6.2
- PGI-Fortran 17.7 combined with GNU-C/C++ 6.3
- PGI 17.7
Known compilers to fail include:
- PGI-Fortran 17.10, 18.1
Required dependencies:
- CMake --- For use and installation see http://www.cmake.org/
- ecbuild --- ECMWF library of CMake macros
- eckit (with MPI support) --- C++ support library
Recommended dependencies:
- fckit --- For enabling Fortran interfaces
- python (only when Fortran bindings are required)
Optional dependencies:
- gridtools-storage --- For GPU interoperability
- transi --- For enabling IFS spherical harmonics transforms ( not open-source )
- CGAL --- For enabling Delaunay triangulation of unstructured grids
- Eigen3 -- For certain linear algebra operations
- FFTW -- For enabling inverse spherical harmonics transforms (TransLocal)
Atlas employs an out-of-source build/install based on CMake.
Make sure ecbuild, eckit and fckit are installed and the ecbuild
executable script is found ( which ecbuild
). Following environment variables
help the build system to detect the right dependencies:
# For finding eckit
ECKIT_PATH # Path to eckit prefix
# For finding fckit
FCKIT_PATH # Path to fckit prefix
Other environment variables that could be required for optional features:
# For finding gridtools-storage
GRIDTOOLS_STORAGE_PATH # Path to gridtools-storage prefix
# For finding transi
TRANSI_PATH # Path to transi prefix
# For finding CGAL
BOOST_ROOT # Path to Boost prefix
CGAL_DIR # Path to directory containing CGALConfig.cmake
Eigen3_DIR # Path to directory containing Eigen3Config.cmake
FFTW_PATH # Path to FFTW prefix
Now proceed with installation as follows
# Environment --- Edit as needed
ATLAS_SRC=$(pwd)
ATLAS_BUILD=build
ATLAS_INSTALL=$HOME/local
# 1. Create the build directory:
mkdir $ATLAS_BUILD
cd $ATLAS_BUILD
# 2. Run CMake
ecbuild --prefix=$ATLAS_INSTALL -- $ATLAS_SRC
# 3. Compile / Install
make -j10
make install
# 4. Check installation
$ATLAS_INSTALL/bin/atlas --info
Extra flags maybe added to step 2 to fine-tune configuration.
--build=DEBUG|RELEASE|BIT
--- Optimisation level- DEBUG: No optimisation (
-O0 -g
) - BIT: Maximum optimisation while remaning bit-reproducible (
-O2 -g
) - RELEASE: Maximum optimisation (
-O3
)
- DEBUG: No optimisation (
-DENABLE_OMP=OFF
--- Disable OpenMP-DENABLE_FORTRAN=OFF
--- Disable Compilation of Fortran bindings
Note: By default compilation is done using shared libraries. Some systems have linking problems with static libraries that have not been compiled with
-fPIC
. In this case, also compile atlas using static linking, by adding to step 2:--static
Atlas behaviour can be configured through some environment variables with defaults marked in square brackets
ATLAS_INFO=<0|[1]>
--- Control printing of Atlas standard informationATLAS_DEBUG=<[0]|1>
--- Control printing of Atlas debug informationATLAS_TRACE=<[0]|1>
--- Control printing of Atlas traces (includes timings)
Contributions to Atlas are welcome. In order to do so, please open an issue where a feature request or bug can be discussed. Then issue a pull request with your contribution. Pull requests must be issued against the develop branch.
If you publish work which mentions Atlas, or Atlas has been useful in your research, please cite the following paper:
@article{DECONINCK2017188,
title = "Atlas : A library for numerical weather prediction and climate modelling",
journal = "Computer Physics Communications",
volume = "220",
pages = "188 - 204",
year = "2017",
issn = "0010-4655",
doi = "https://doi.org/10.1016/j.cpc.2017.07.006",
url = "http://www.sciencedirect.com/science/article/pii/S0010465517302138",
author = "Willem Deconinck and Peter Bauer and Michail Diamantakis and Mats Hamrud and Christian KΓΌhnlein and Pedro Maciel and Gianmarco Mengaldo and Tiago Quintino and Baudouin Raoult and Piotr K. Smolarkiewicz and Nils P. Wedi",
keywords = "Numerical weather prediction, Climate, Earth system, High performance computing, Meteorology, Flexible mesh data structure"
}