Skip to content

ALaDyn/psc

 
 

Repository files navigation

PSC

Documentation Status Build Status

Introduction

This is the development version of PSC. It is usable for certain problems, but overall lots of changes are happening and things may be unstable / evolving.

PSC is a 3-dimensional fully electromagnetic particle-in-cell code for kinetic plasma simulations. It supports Nvidia GPUs and patch-based dynamic load balancing.

Quickstart

Building the code

PSC now uses cmake as its buildsystem. It is strongly recommended to build the code outside of the source in a different build directory. This is what I usually do:

$ cd src/psc # This is where the cloned git repository is (ie, the source)
$ mkdir build # This is my build directory
$ # create a cmake.sh so I don't have to remember the options I used
$ cat > cmake.sh <<EOF
cmake \
    -DCMAKE_BUILD_TYPE=Release \
    -DUSE_CUDA=OFF \
    -DUSE_VPIC=OFF \
    ..
EOF
$ . cmake.sh # run cmake (hope for the best)
$ make

The above should build a lot of tests (you can use ctest . to run them) and two actual cases, psc_bubble_yz and psc_flatfoil_yz which you'll find in the src/ subdirectory inside your build directory.

Running a case

Make a directory for your run and change into it. This generation of PSC uses (almost) no command line options, rather everything gets hardcoded into a case (a.k.a. "input deck" in VPIC speak), e.g., psc_flatfoil_yz.cxx. (Eventually, the build system should be set up where a standalone case can be easily compiled, rather than expecting for it to be in the PSC source directory). So it should be as easy as

mpirun -n 4 path/to/build/src/psc_flatfoil_yz

if you don't have to go through a batch system.

Documentation

Well, this is really a big todo, so for now a lot of it is probably all about emailing kai.germaschewski@unh.edu. But I'm working on actually usable documentation and there's a start here: https://psc.readthedocs.io/en/latest/

About

The PSC particle-in-cell code

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C 48.4%
  • C++ 31.6%
  • Cuda 5.6%
  • Fortran 4.6%
  • Shell 3.7%
  • Python 2.2%
  • Other 3.9%