[saeed jamshidi]

Hi all,

I going to run my simulation, but there is not enough space on my computer. Is there any alternative to come up with?
Some space is occupied by the parallel running and other because of reconstruction!

[Yann]

Hello Saeed,

Reconstructing data is unnecessary. You can visualize decomposed data with ParaView native reader for OpenFOAM. Instead of using paraFoam, use paraFoam -builtin or just create an empty file with .foam extension (for instance touch case.foam) and load it in paraview. You will have an option to select if you want to load the reconstructed or decomposed case.

Hope this helps,
Yann

[saeed jamshidi]

Hi Yann,
Thanks for your reply.

So you are saying I should follow below procedures:

cd ${0%/*} || exit 1
clear
source clean
source .color
#-----------------------------------------------------------------------
echo -e "   - Copy 0.orig to 0"
cp -r 0.orig 0
#------------------------------------------------------------------------------
echo -e "   - Create background mesh"
blockMesh
surfaceFeatureExtract
#------------------------------------------------------------------------------
echo -e "   - Meshing with snappyHexMesh"
mpirun -np 6 snappyHexMesh -overwrite 
checkMesh
#-----------------------------------------------------------------------
echo -e "   - Decomposition"
decomposePar
#-----------------------------------------------------------------------
echo -e "   - run"
mpirun -np 6 pimpleFoam -parallel

#-----------------------------------------------------------------------
echo -e "   - reconstruction"
 #   reconstructPar //According Yann opinion
#-----------------------------------------------------------------------
echo -e "   - Demonstration"
postProcess -func vorticity
touch foam.foam
paraview foam.foam

#------------------------------------------------------------------------------

[Yann]

Hello Saeed,

Yes, except your should run the postProcess command in parallel:

mpirun -np 6 postProcess -func vorticity -parallel

Once you have launched paraview, don't forget to switch the case type from "reconstructed case" to "decomposed case".

Cheers,
Yann

[saeed jamshidi]

Thank you Yann,

Does write format of binary instead of ascci in controlDict reduce size of exported files in zero folder?

[Yann]

Yes, using binary format should use less space on the time directories written to disk.

[saeed jamshidi]

Hi again,

Just ending this thread, I also found that when we use cache utility in fvsolution in order to save previous time step date and accelerate running, it occupies significant space of your PC. So, we can pause running and resume it again for the storage space recovery.

cache
{
    grad(U);
}

[saeed jamshidi]

Hi again,


In three dimensional simulation due to large number of cells we may encounter with lack of storage space! Since we write various parameters such as U, P, T and ... and they have large size.

The question: is it possible to store run data on other drives of our computer instead of 'C'?!

[Yann]

Hello,

You can run from any place, as long as the computers running OpenFOAM have permissions to access the case files.

Performance may vary depending on read/write times on the storage you are using.

Regards,
Yann

[saeed jamshidi]

Hi Yann,

thank you for the reply.

Could you give me more assistance?

According to the picture I am using Ubuntu. It directly stores data on C.

[Yann]

Alright, you're using WSL.
Your drives should be listed in /mnt/ so you should be able to access it doing something like this for instance:

cd /mnt/e/

Internal drives should be mounted by default. If you want to use an external drive you may have to mount it manually

Cheers,
Yann

[saeed jamshidi]

Thank you Yann,
it worked for me.