# Fortran-only Before the python bindings were introduced, GVEC was Fortran-only. Here, installation and use of the Fortran-only version is described, which might be helpful from a developers' perspective. ## Install Fortran executable of GVEC with `cmake` ```{warning} This will only install the Fortran executables of GVEC, which are mostly used for testing. The python bindings are not installed with this option. ``` The standard way of compiling GVEC is using cmake presets, but there is also an interactive way with ccmake. :::::{note} Before executing cmake, be sure that you have all libraries (netcdf must be compiled in serial). It might also be necessary to export an environment variable `FC` to point to the compiler. ::::{tab-set} :sync-group: compiler :::{tab-item} GNU :sync: gnu ```bash export FC=`which gfortran` ``` ::: :::{tab-item} Intel ifx :sync: intel ```bash export FC=`which ifx` ``` ::: :::{tab-item} Intel ifort :sync: intel ```bash export FC=`which ifort` ``` ::: :::: ::::: ::::::{tab-set} :::::{tab-item} CMake Presets With Cmake version > 3.22, the CMakePresets feature can be used to configure and then build the code. 1. Start from the GVEC directory with ```bash cmake --list-presets ``` to show a list of presets (defined `CMakePresets.json` and `CMakeUserPresets.json`). 1. Select a preset and specify the `build` directory (the build directory can have any name). ::::{tab-set} :sync-group: os :::{tab-item} Linux :sync: linux ```bash cmake --preset gvec_config_release -B build ``` ::: :::{tab-item} MacOS :sync: mac ```bash cmake --preset gvec_config_release_mac_brew -B build ``` ::: :::: 1. Then compile with (`-j` compiles in parallel) ```bash cmake --build build -j ``` Further, the user can also create own presets by creating his own preset file `CMakeUserPresets.json` in the GVEC directory. Be careful to only add new entries with new names, as they must be different from those in `CMakePresets.json`. For example compiling on a mac in debug mode: ```json { "version": 3, "cmakeMinimumRequired": { "major": 3, "minor": 22, "patch": 0 }, "configurePresets": [ { "name": "gvec_config_debug_mac", "displayName": "GVEC configure: default debug build on a MAC", "hidden": false, "cacheVariables": { "CMAKE_BUILD_TYPE": "Debug", "COMPILE_GVEC": "On", "CMAKE_HOSTNAME": "mac_brew", "LINK_GVEC_TO_NETCDF": "On", "USE_OPENMP": "On", "COMPILE_GVEC_AS_STATIC_LIB": "On" } } ] } ``` The user presets then appear also on the list of presets. :::{note} The preset files allow building the code in **VScode** with "CMake" and "CMake Tools" extensions. ::: ::::: :::::{tab-item} CCMake Interactive To compile GVEC interactively (needs `ccmake` command): 1. create a new subdirectory that can have any name, e.g. `build` ```bash mkdir build ; cd build ``` 1. Inside that directory execute ```bash ccmake ../ ``` `ccmake` gives you a visual setup on the terminal. * Press "enter" to change options, and press "enter" again to fix the change * Press "c" to configure and "g" to create the Makefiles. * If `BUILD_NETCDF=ON` and no preinstalled libraries for netcdf are found, an error occurs... * On a Mac, be sure to activate `COMPILE_GVEC_AS_STATIC_LIB=ON` (in ccmake, toggle to view all variables by typing `t`) * In the main `CMakeList.txt` file, some pre-defined setups (library paths) for different architectures are controlled by setting the `CMAKE_HOSTNAME` to `cobra`/`raven`/`mac_brew`/`mac_ports`/`tokp`/.. . 1. Finally, compile GVEC in the build directory by typing (`-j` compiles in parallel) ```bash make -j ``` ::::: :::::: Now GVEC should be installed! You should find the `gvec` and `gvec_post` binaries in `build/bin/` and can continue with [](/tutorials/getting-started). ## Run the GVEC Fortran executable ```{warning} For users, we suggest using pygvec instead of the GVEC Fortran executable! It has has a simpler parameterfile (ending with `.ini`) and thus less features than using `pygvec` with TOML/YAML parameterfiles. It is mainly used for testing purposes. ``` 1) To install the Fortran executable of GVEC, follow the [installation instructions](/user/install). 2) The binary executables `gvec` and `gvec_post` should now be found in `build/bin/`. 3) GVEC is configured with a custom parameter file, typically called `parameter.ini`. Example parameter files are found in `ini/` or `test-CI/examples/` ### Running GVEC There are several test example input files named `parameter.ini`, which are found in a subfolder of [`test-CI/examples` {fab}`square-gitlab`](https://gitlab.mpcdf.mpg.de/gvec-group/gvec/-/blob/develop/test-CI/examples/). * For execution, go into one of these folders and execute for example the following commands ```bash cd test-CI/examples/ellipstell_lowres ../../../build/bin/gvec parameter.ini |tee log # (|tee pipes the screen output also into the file `log`) ``` * You can also restart a simulation by using one of the restart files (`*_State_*.dat`). Before the restart, resolution parameters in the `.ini` file can be changed, so that the new iterations will be on a finer grid, for example, or with more modes. The restart is triggered by simply adding the restart filename as an argument to the execution command, for example: ```bash ../../build/bin/gvec parameter.ini ELLIPSTELL_State_0000_00000200.dat |tee log ``` Then the first integer (`_0000_`) will be incremented for the newly written restart files. #### Run GVEC with OpenMP If you run gvec with the OpenMP parallelization, be sure to set the desired number of threads as an environment variable: ```bash #replace ??? by the number of threads you want to use export OMP_NUM_THREADS=??? ``` ### Running tests After compilation, you can quickly run some tests via `ctest`, that then calls the `pytest` environment of GVEC (requires `python >3.10` to be installed!). Change to the build directory, and execute: ```bash ctest -T test --output-on-failure -R ``` ### Visualization Using the python interface, any statefile can be loaded and visualized using the `ipython` notebook [`visu.ipynb`]() (view [online {fab}`square-gitlab`](https://gitlab.mpcdf.mpg.de/gvec-group/gvec/-/blob/develop/python/examples/visu.ipynb)). For line plots, csv datafiles are generated. For 3D visualization data, it is possible to write `*visu*.vtu` files, that can be visualized in [paraview](https://www.paraview.org). There is an option to write visualization data in netcdf, `*visu*.nc`, which can be read for example in python.