View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0002187 | OpenFOAM | Bug | public | 2016-08-11 14:43 | 2016-09-08 11:25 |
Reporter | r0383899 | Assigned To | henry | ||
Priority | normal | Severity | minor | Reproducibility | always |
Status | closed | Resolution | no change required | ||
Platform | GNU/Linux | OS | Ubuntu | OS Version | 15.04 |
Summary | 0002187: combination multiLevel + singleProcessorFaceSets constraint doesn't behave as expected. | ||||
Description | When decomposing a mesh using the multiLevel method with two levels, I tried to impose the singleProcessorFaceSets constraint on the first level only by placing it in the level0 dictionary. However, when running decomposePar this way, the constraint is neglected. The constraint is only taken into account when it is defined on top of the decomposeParDict file and not inside the multiLevelCoeffs dictionary. Am I defining it incorrectly or is this currently not supported by OF? If not, is there perhaps a workaround? | ||||
Tags | No tags attached. | ||||
|
decomposeParDict (1,200 bytes)
/*--------------------------------*- C++ -*----------------------------------*\ | ========= | | | \\ / F ield | OpenFOAM: The Open Source CFD Toolbox | | \\ / O peration | Version: 3.0.x | | \\ / A nd | Web: www.OpenFOAM.org | | \\/ M anipulation | | \*---------------------------------------------------------------------------*/ FoamFile { version 2.0; format ascii; class dictionary; note "mesh decomposition control dictionary"; object decomposeParDict; } // * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * // numberOfSubdomains 48; method multiLevel; multiLevelCoeffs { level0 { numberOfSubdomains 4; singleProcessorFaceSets ((f0 -1)); method scotch; } level1 { numberOfSubdomains 12; method scotch; } } // ************************************************************************* // |
|
This is currently not supported. At the lower levels there is no knowledge of the mesh structure (e.g. special patch types); it works on a graph representation only. As a workaround maybe you can use a hierarchical decomposition at the lower levels and play with the processor weights to get your problem areas onto a single processor. |
Date Modified | Username | Field | Change |
---|---|---|---|
2016-08-11 14:43 | r0383899 | New Issue | |
2016-08-11 14:43 | r0383899 | File Added: decomposeParDict | |
2016-09-08 11:22 | MattijsJ | Note Added: 0006850 | |
2016-09-08 11:25 | henry | Assigned To | => henry |
2016-09-08 11:25 | henry | Status | new => closed |
2016-09-08 11:25 | henry | Resolution | open => no change required |