Jan Totz, jantotz@itp.tu-berlin.de
- Linux Ubuntu >= 18.04
- Boost >= 1.67
- CUDA >= 9.2
- python >= 3.6
- GNU make >= 4.1
- gcc >= 7.3.0
- download the latest version from: http://www.boost.org/users/download/
- unzip tar.bz2 in $HOME/source:
tar xvjf ~/Downloads/boost*.tar.bz2 -C ~/source
cd ~/source/boost*
./bootstrap.sh
sudo ./b2 install
(only required if you want to use GPU-accelerated code, otherwise skip to 3)
- get .run file(s) (possible patches) from cuda repository https://developer.nvidia.com/cuda-toolkit
- save both in ~/cuda
chmod u+x ~/cuda/*
sudo service lightdm stop
sudo ./cuda_*.run --silent --override --toolkit --samples
sudo apt-get install python3 python3-pip
sudo -H pip3 install --upgrade pip
sudo -H pip3 install jupyter numpy matplotlib
for the CPU version:
cd coupled_oscillators_solver
# adapt line 13 in the makefile for CPU or GPU version by commenting it
make
- modify .ini file in ini directory to your liking
- run program via
./coupled_oscillators_solver.exe --ini ../ini/<nameOfIni>.ini
- for example:
./coupled_oscillators_solver.exe --ini ../ini/phase_global.ini
jupyter notebook
and navigate to the directory with the .ipynb file
- activate cells with shift+enter
- enter name and continue