[AndreasPe]

Hello,

i am curious if somebody knows if the collated file format which was introduced in OpenFOAM v1712 and OpenFOAM 5.0 will also be integrated into the Foam-Extend project?

I also have a question regarding the file handler itself: Why doesn't the collated file format directly write the data in the "reconstructed" way instead of having each processor field in a seperate region of the collated files? Is it because of the way how the mesh is handled with parallel processing? I will have a look into the commit (e.g. https://github.com/OpenFOAM/OpenFOAM...ae435d4ab42e6f ) but any help would be greatly appreciated.

Thank you for your help!

Andreas

[hjasak]

Hi Andreas,

I have no problem including the collated file format in foam-extend. This seems to belong into a bigger project we are working on which includes low-level support for load balancing and alternative ways of dealing with the parallel decomposition mesh numbering.

Would you be interested in preparing a git commit for the merge?

Hrvoje

[iLyrics]

Hi Hrvoje, the current dynamic load balancing tools aren't easy to use in our calculations when it is combined with adaptive mesh refinement. How is the work you mentioned going?

[sphani4391]

Originally Posted by hjasak

Hi Andreas,

I have no problem including the collated file format in foam-extend. This seems to belong into a bigger project we are working on which includes low-level support for load balancing and alternative ways of dealing with the parallel decomposition mesh numbering.

Would you be interested in preparing a git commit for the merge?

Hrvoje

Hi Prof. Jasak,


I am using fe32 to run a large transient simulation in a cluster. Due to the number of files limitation, I am unable to write much timesteps. A collated fileHandler is really needed for me to ensure I don't generate a lot of files. Is this feature available in fe32?