decomposePar not decomposing refinement history

kmooney · July 8, 2010, 14:23

Hi all,

I have been getting remeshing inconsistancies between serial and parallel runs when using interDyMFoam (1.6.x and 1.7.x) and think that I have figured out the issue.

I have been using interDyMFoam in serial as somewhat of a preprocessor/meshing tool to get a fine mesh around my interface. Once I had the mesh built I decomposed it and attempted to run it in parallel only to have dynamicRefineMesh re-refine the cells that were already adequate. It seems that the refinementHistory file is not being decomposed and passed into the processorX folders.

Is there a way to readily decompose that file?

mattijs · July 9, 2010, 05:01

Correct, the refinementHistory is not decomposed.

(the problem with it is that the history can refer to local cells only. So a valid decomposition would have to make sure that it keeps a cluster of cells originating from the same cell together. In other words base the decomposition on the coarsest level (= starting level). Not easy and then one might as well just decompose the starting mesh and preprocess in parallel)

Mattijs

kmooney · July 9, 2010, 09:48

Thanks for the reply!

It does seem that that would be a non-trivial problem. Preprocessing in parallel seems to be working. Thanks a lot.

Kyle

July 8, 2010, 14:23	decomposePar not decomposing refinement history	#1
kmooney Senior Member Kyle Mooney Join Date: Jul 2009 Location: San Francisco, CA USA Posts: 323 Rep Power: 17	Hi all, I have been getting remeshing inconsistancies between serial and parallel runs when using interDyMFoam (1.6.x and 1.7.x) and think that I have figured out the issue. I have been using interDyMFoam in serial as somewhat of a preprocessor/meshing tool to get a fine mesh around my interface. Once I had the mesh built I decomposed it and attempted to run it in parallel only to have dynamicRefineMesh re-refine the cells that were already adequate. It seems that the refinementHistory file is not being decomposed and passed into the processorX folders. Is there a way to readily decompose that file?

July 9, 2010, 09:48		#3
kmooney Senior Member Kyle Mooney Join Date: Jul 2009 Location: San Francisco, CA USA Posts: 323 Rep Power: 17	Thanks for the reply! It does seem that that would be a non-trivial problem. Preprocessing in parallel seems to be working. Thanks a lot. Kyle Minada likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[snappyHexMesh] snappyHexMesh not refining surfaces	Hydro1004	OpenFOAM Meshing & Mesh Conversion	3	August 29, 2012 11:56
DecomposePar unequal number of shared faces	maka	OpenFOAM Pre-Processing	6	August 12, 2010 09:01
[snappyHexMesh] external flow with snappyHexMesh	chelvistero	OpenFOAM Meshing & Mesh Conversion	11	January 15, 2010 19:43
decomposePar hierarchicalCoeffs	maka	OpenFOAM Running, Solving & CFD	1	August 14, 2008 17:39
Mesh Refinement	JY	Siemens	7	September 19, 2002 13:37

July 9, 2010, 05:01		#2
mattijs Senior Member Mattijs Janssens Join Date: Mar 2009 Posts: 1,419 Rep Power: 26	Correct, the refinementHistory is not decomposed. (the problem with it is that the history can refer to local cells only. So a valid decomposition would have to make sure that it keeps a cluster of cells originating from the same cell together. In other words base the decomposition on the coarsest level (= starting level). Not easy and then one might as well just decompose the starting mesh and preprocess in parallel) Mattijs