Change mesh location on drive..
Hi..
I have a mesh, which is want to do 12 simulations on (changing inflow direction).. The mesh I have right now is app 7.5m cells, which is about 1.5GB of data.. If I want to run these simultanious, I will need to generate 12 cases, each with a copy of the mesh.. It'll take up a lot of space copying the grid to each sector (14.5GB of duplicate data), but worse, it takes time.. I was wondering if there's a way of changing the reference to the mesh location, so that the simulations can share the same mesh.. Cheers Mark Herskind |
Greetings Mark,
It depends on where the mesh is kept at. For example, when we run "blockMesh", the mesh is kept in the folder "constant/polyMesh". The same happens when we run "snappyHexMesh -overwrite", which will overwrite the latest mesh (which is usually the initial mesh by blockMesh). So, if you can keep the mesh in the folder "constant/polyMesh" then you can use symbolic links for sharing the mesh between cases. For example, if we have 3 case folders: "case1 case2 case3". Generate the mesh in case1. Then step back into the previous folder and run: Code:
rm -rf case2/constant/polyMesh Code:
cd case3/constant But the best method would be to use absolute folders, as long as you run always on the same path: Code:
ls -s $PWD/case1/constant/polyMesh case2/constant/polyMesh Code:
echo $PWD Bruno |
All times are GMT -4. The time now is 11:22. |