C7. Sod shock tube in rotated channel with AMR
Case description
This is the standard Sod shock tube problem. The geometry is a rectangular channel at \(30^\circ\) to the mesh. This case is run with multiple levels of AMR. The profiles shown below are through the channel centerline.
Density at t=0.1s

Field profiles in the centerline at t=0.1s




Running study
paren=`pwd`
pelec="${paren}/PeleC3d.gnu.MPI.ex"
mpi_ranks=36
lev=( 0 1 2 )
for i in "${lev[@]}"
do
rm -rf "${i}"
mkdir "${i}"
cd "${i}" || exit
cp "${paren}/inputs_3d" .
srun -n ${mpi_ranks} "${pelec}" inputs_3d `python3 ${paren}/gen_tube_input.py -a 30.0 -n 8` amr.max_level="${i}" > out
ls -1v *plt*/Header | tee movie.visit
cd "${paren}" || exit
done