A C++ library for developing parallel simulations on Cartesian grids
Schnek is a new library for C++ that makes it easy to develop large parallel simulation codes. The library is intended mainly for use in physics simulations on regular grids but some of its features might be useful for other types of simulation codes.
When writing a new simulation code from scratch one is confronted with a choice. One can either keep the code simple and focus on the physics and the numerical side. For this kind of code, most time is spent on actually developing the interesting stuff. The downside of this type of code is that is not much use to anybody else except the developer of the code herself and maybe the immediate collaborators in a research group.
The other option is to develop a code that might be useful for others, that can be configured using a configuration file and that is written to be versatile and adaptable to problems other than the one it was initially developed for. The code should be easy to use by others and hopefully should also be extendable by others. Writing this type of code requires a lot of time and effort. Only very few projects get to a stage where this investment of time and work is worthwhile.
In addition, if the code handles large amounts of data and should be able to run on multi-processor clusters then additional code has to be written. For simulations on regular grids this means that one has to use multi-dimensional arrays, define boundary values, synchronise grid values accross multiple processes and write data into files that can be used by other software.
Schnek provides a single library that aids in all these tasks and thus massively reduces the code that needs to be written by the developer.
Features of Schnek are
- setup file reader with C style syntax
- extendable for formulas in setup files
- hierarchical structure of simulation components
- mutli-dimensional grids
- initialisation of grid from formulas in setup file
- boundary cells and ghost cells
- MPI support for parallelisation
- MPI update of ghost cells to neighbouring processes
- easy to use parallel and serial HDF5 output of grids
- literature reference system to encourage collaborative development of large codes
Documentation and tutorials can be found on the Schnek website at
Step 1: Download
You can obtain the latest release of Schnek on the GitHub Releases page, (https://github.com/holgerschmitz/Schnek/releases). This should be relatively bug-free. If you want to use Schnek for your own projects then this is your choice.
If you are interested in contributing and helping to improve Schnek, you are very welcome to do so. In this case you can grab the latest development version using Subversion. The command is
git clone https://github.com/holgerschmitz/Schnek.git
Step 2: Configuration
The configuration and installation script have been generated using the GNU automake utility. For the typical configuration simply type
in the directory that you unpacked Schnek into. This will run the configure script and create make files for the next step of the installation process.
For more control over the configuration you can pass options to configure. If you want to install Schnek into a directory different from the standard installation directory, use the --prefix option, like this
Schnek can be compiled with or without MPI support. By default Schnek will be compiled with MPI support if MPI can be found. You can force MPI to be enabled or disabled using the --with-mpi option. The following command will disable MPI even on those systems where it is present.
If MPI is enabled, Schnek will be compiled using the mpic++ compiler command. If you want to control mpic++ you should set your environment variables according to the documentation of your MPI version.
You can switch HDF5 support on or off using the --with-hdf5 option. You can specify yes or no to enable or disable HDF5. By default, HDF5 support is enabled. If your HDF5 headers are located in a non-standard location you can also specify the path to your HDF5 installation.
A full list of available options written out with the --help option.
Step 3: Building the library
Once you have configured Schnek, type
to build the library.
Step 4: Installing
Finally, install the library by typing
sudo make install
depending on the installation directory.
Step 5: Include Schnek in your projects
By default the Schnek headers are located in /usr/local/include/schnek/. To include the headers in your C++ files, you should include them like this.
You nee to link your code to the Schnek libraries. For most compilers this should be achieved by adding -lschnek to your linker flags.
License and Disclaimer
Schnek is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
Schnek is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with Schnek. If not, see http://www.gnu.org/licenses/.