Cubit is unresponsive after meshing is completed

Coreform Cubit General Topics
#1

Hi,

after meshing a complex geometry, the program is unresponsive for a longer period of time (usually much longer than the actual meshing time). The output in the console is shown below:

WARNING: >>>>Poor Quality Shear Hex Generated!<<<< on Volume 148
    (For example, the Shear metric for Hex 200610 is 0.0605919 .)
    The threshold for a Hex is 0.2

Generated 38474 hexes
Meshing time: 12.664080 

Finished Command: mesh volume 6 to 19 51 to 53 55 to 67 116 to 277

Afterwards, it becomes responsive again without any additional output.

I am using version 2022.11 on Kubuntu 22.04 with a floating commercial license which is behind a VPN. Are there additionally parameters (e.g. a debug mode), which I could set to find out what is causing the hang-ups? Is there an additional optimization step, which is running after the meshing?

Thanks!

Best,
Paul

#2

Can you paste the full command line output please? There was a bug in development versions of Cubit that resembled this, but we thought we made sure that it didn’t make it out to any customers in that release.

You could also try the latest Development Build of Cubit available from Coreform Product Downloads, we have made a fix for the aforementioned bug.

#3

Attached is the full console output for the meshing command using Coreform-Cubit-2023.1-rc+33337.

console_output.txt (4.0 KB)

The hang-up is between the last two outputs:

Finished Command: mesh volume 35 66 67 81 to 93 96 to 102 105 to 109 113 114 116 to 169 172 to 174 176 180 184

Finished Command: undo group end

Besides the warnings about mesh quality, I noticed there is an error right at the beginning:

ERROR: invalid syntax in parse_cubit_list
#4

I believe that is a different error than what I was talking about, I’ll pass this on as a separate bug.

#5

Could you attach a journal file that replicates the issue please?

#6

Unfortunately, it uses an imported geometry from a step-file, which I cannot share and I do not have a minimum working example. Is there anything I can do on my side?

#7

The journal file could still be useful. I am interested in seeing if I can track down the error with parse_cubit_list.

One thing you might try is doing an undo off at the beginning of the journal file. That way you won’t do incremental saves to disk which can be expensive.

Thanks,
Karl

#8

I fear that it might not be very useful as well, since I do not have a continuous journal file as I have been trying to decompose the geometry in multiple tries due to the complexity:

history09.jou (15.8 KB)

Disabling the undo functionality did not help.

#9

Hi,

It looks like you are trying to do a hex mesh, right? All I see are setting volumes to scheme auto. There may be a surface that is causing problems. The paver is pretty good about quitting when it can’t complete the algorithm correctly. You might try meshing groups of surfaces and see if that helps identify the problem area.

Thanks,
Karl

#10

Yes, I am trying to build a full hex mesh. I tried meshing the group of surfaces, which seems to work:

Matching intervals successful.
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh
Smoothing Surface Mesh

WARNING: >>>>Poor Quality Shape Face Generated!<<<< on Surface 2777
    (For example, the Shape metric for Face 376546 is 0.192933 .)
    The threshold for a Face is 0.2

Smoothing Surface Mesh

WARNING: >>>>Poor Quality Shape Face Generated!<<<< on Surface 2785
    (For example, the Shape metric for Face 377047 is 0.122622 .)
    The threshold for a Face is 0.2

Smoothing Surface Mesh

WARNING: >>>>Poor Quality Shape Face Generated!<<<< on Surface 2848
    (For example, the Shape metric for Face 377588 is 0.0856395 .)
    The threshold for a Face is 0.2

Smoothing Surface Mesh
Smoothing Surface Mesh
Surfaces 1350, 1351, 1360, 1361, 1370, 1371, 1380, 1381, 1390, 1391, 1400, 1401,
   1410, 1411, 1420, 1421, 1431, 1440, 1441, 1451, 1460, 1461, 1471, 1480, 1481,
   1490, 1491, 1517, 2466, 2618, 2630, 2743, 2750, 2777, 2785, 2848, 2853, 2858 meshing completed using scheme: pave
Surfaces 426 meshing completed using scheme: map
Generated 5968 quads
Meshing time: 1.005806 

Finished Command: mesh surface 426 1350 1351 1360 1361 1370 1371 1380 1381 1390 1391 1400 1401 1410 1411 1420 1421 1431 1440 1441 1451 1460 1461 1471 1480 1481 1490 1491 1517 2466 2618 2630 2743 2750 2777 2785 2848 2853 2858

Finished Command: undo group end

After the mesh is completed, cubit hangs for at least a few seconds after the Finished Command: mesh surface .... It is not as bad as in the 3D case, where the hang-ups were over a minute. When I try to mesh the volume afterwards, it aborts immediately. I guess at least this way I can avoid the waiting time when directly meshing the volume and keep working on the decomposition.

#11

An update: While working now more with tetmesh as the geometries got too complex, I noticed that the hang-ups happen there as well but only if I try to mesh the geometry using the GUI in the following way:

Mode - Mesh → Entity Volumes → Action Intervals → Automatic Sizing → Mesh

The options for tetmesh have already been set at this point, and it meshes but I have to terminate Cubit as it becomes unresponsive:

ERROR: invalid syntax in parse_cubit_list                                                                                                                        
Mesh deleted and propagated on:
Volume 1
Finished Command: delete mesh volume all propagate

Finished Command: undo group begin

Auto-sizing requested with sizing factor of 8.000000
Finished Command: volume all size auto factor 8

Matching intervals successful.
Begin tri meshing 1918 surfaces 1 to 1918
Generating 1-th hyperpatch mesh
Generating 3-th hyperpatch mesh
Generating 4-th hyperpatch mesh
Generating 2-th hyperpatch mesh
Generating 5-th hyperpatch mesh
Generating 6-th hyperpatch mesh
Generating 7-th hyperpatch mesh
Generating 8-th hyperpatch mesh
Generating 9-th hyperpatch mesh
Generating 10-th hyperpatch mesh
Completed trimeshing 1918 surfaces

WARNING: >>>>Poor Quality Shape Tet Generated!<<<< on Volume 1
    (For example, the Shape metric for Tet 580359 is 0.104158 .)
    The threshold for a Tet is 0.2

Mesh indicates volume 1 (Volume 1) has 222+ through holes.
Volume 1 meshing completed using scheme: tetmesh
Generated 166883 tets
Meshing time: 12.981370 

Finished Command: mesh volume all

Again there is the ERROR: invalid syntax in parse_cubit_list at the beginning. However, if I go through the following steps, there is no error and Cubit does not hang:

Mode - Mesh → Entity Volumes → Action Mesh → Tetmesh → Mesh

and below the following output:

Finished Command: volume 1 scheme tetmesh

Matching intervals successful.
Begin tri meshing 1918 surfaces 1 to 1908, 1910 to 1918
Completed trimeshing 1917 surfaces

WARNING: >>>>Poor Quality Shape Tet Generated!<<<< on Volume 1
    (For example, the Shape metric for Tet 567292 is 0.1087 .)
    The threshold for a Tet is 0.2

Mesh indicates volume 1 (Volume 1) has 236+ through holes.
Volume 1 meshing completed using scheme: tetmesh
Generated 164025 tets
Meshing time: 22.474350 

Finished Command: mesh volume 1

So maybe the problem with the hang-ups is due to triggering the meshing process from Action Intervals -> Automatic Sizing?

#12

Hi @PaulOffTheWall,

Cubit should not be aborting (i.e. crashing). If you can give us a sample problem, we should fix that! We have recently made a change that should speed up sizing dramatically if that is the cause of the slowness. That change is in latest development release. That release should be in pretty good shape right now. Coreform Product Downloads

Karl