[nico7]

Hello,

I have access to a quad cpu (opteron 6276, 16 cores per cpu) dell machine. I've been looking at parallel efficiency running the LTSInterFoam solver with the wigley tutorial.

First results I got is shown in the following graph; this is running the small (0.2M cells) mesh (as defined in the tutorial)


I've then refine the mesh (by refining blockmesh and surface mesh of shm); I get the following results:


Not looking that great.
Should I expect something different or similar for the kind of hardward I am using?
I've run parallel jobs using the runParallel command so no options for mpirun; anything particular that I should use?
Compilation was also basic one (with openmpi).

Thanks for reading,
Nico