This post will walk you through the procedure to build and test ChaNGa over Charm++ using UCX, or Open MPI/Intel MPI machine layers over InfiniBand Interconnect and Intel architecture (Skylake).

...

Now after successful build and test of charm++ you are ready for ChaNGa build.

Build ChaNGa

Before you start, set the global variable for the charm directory.

...

Note: if you compile several times with different charm++ builds, make sure to clean the old compilations to leave no tracks.

Code Block
$ make clean && make ref-clean && make dist-clean ...

Configure,

Code Block
$ CHARMC=$PWD/../configure .../charm/mpi-linux-x86_64-ifort-icc/bin/charmc CC=icc CXX=icpc FC=ifort ./configure

Build,

Code Block
$ make -j16

...

In this example, we will use small dwarf star dwf1 (dwf1.2048.param and dwf1.2048.00384) is a 5 million particle zoom-in simulation. It is cosmological, but the particle sampling focuses on a single halo of roughly 1e11 solar masses. This is a dark matter only version of the DWF1 model that demonstrates how disk galaxies can form in a cosmological context.

To test ChaNGa, download dwf files into ChaNGa directory (where the executable is).

Code Block
$ ls dwf1.2048.* dwf1.2048.00384 dwf1.2048.param

...