Ligh7bringer / Nbody-parallel

A parallel n-body simulation with CUDA, OpenMP and MPI.

Parallelisation of a brute force n-body simulation O(n^2)

An investigation of possible ways of improving the performance of an O(n^2) n-body simulation. Techniques implemented:

• GPGPU parallelisation with CUDA
• Distributed parallelisation with MPI
• Multithreaded parallelisation with OpenMP

Dependencies

• SFML (included as a submodule)
• OpenMP (supported by most c++ compilers)
• MPI (Windows, on Linux use OpenMPI/mpich)
• Cuda (v.10)

Build

1. Clone this repo.
2. In a terminal run git submodule update --init --recursive
3. Use CMake to generate a solution or makefiles, e.g. cmake -G "Visual Studio 15/Unix Makefiles/etc."
4. Compile from within your IDE of choice or via a terminal using make