And here is a better resolved version with auto scaling turned off.
Here is a short test video of diffusion using PSE in 2D. The top plot shows the concentration field while the bottom is the concentration delta for each time step.
This should be pretty straightforward.
Assuming you already have gcc you need:
Gfortran has an os x installer so it is a breeze to install. After that, build openmpi with the usual:
$ configure —prefix=/Users/milan/Work/Library/openmpi
$ make
$ make install
mpi is big so that takes a while… I install it in a different location because it can have conflicts with the default openmpi that comes with osx (/usr/local didn’t work for me). You can pass the path to mpicc and mpif90 directly to ppm and it will work perfectly.
To compile libppm and libppmnumerics use the following procedure:
$ ./configure LDFLAGS=-L/Users/milan/Work/ETH/Thesis/ppm/metis/ —enable-mpi —with-ppm=../libppm FC=/Users/milan/Work/Library/openmpi/bin/mpif90 CC=/Users/milan/Work/Library/openmpi/bin/mpicc CXX=/Users/milan/Work/Library/openmpi/bin/mpic++
$ make
$ make install
The option “—with-ppm=…” should only be used for libppmnumerics. This should be all that you need to do.
You probably want to test the installation now with the provided example client. You can download it on this page.
I just started preparing for my thesis work in Computer Science at ETH Zurich, and I am supposed to do a simulation of cell membranes using particle methods. The purpose of this blog is to document the research process and, hopefully, to produce some useful or interesting texts on the whole experience. The majority of stuff will be quite technical so it is not for the math impaired. But I will probably do a few posts on the personal experience which could be interesting to random people.
So… Welcome! ;)
