Problems with parallelization in WIND (branch Damping)
I have something I really do not understand and never seen before:
in branch Damping I have changed the recipe for damping in sideEuler.cpp basically DRMIN and DRMAX are computed with a new formula
I have made my run with 1 MPI process , regolded since changing DRMIN and DRMAX affects all the fields and tested with 2,3,4,5 MPI process the test pass with 2 and 3, fails with 4 and 5
I report the plot for gasvrad2.diff the sequence from top left to bottom right is for 2,3,4,5
data are on /beegfs/home/elena/fargOCANew/fargOCA/build/test/integration/3D/WIND/out1 out2 etc..
test with 2 3 and 5 MPI have been done with the command: srun --nodes=1 -p fdr --exclusive runMPI.sh -n
1 and 4 with ctest -R integ3DWind1 integ3DWind4
HELP!