Topic: Crash of a program

Hi Mikael,

I have some problem with input file. It crashes after I run it.
Can you help?

Post's attachments

DSlab.in 583.59 kb, 8 downloads since 2015-03-26 

You don't have the permssions to download the attachments of this post.

Re: Crash of a program

Running it in debug mode I find that

_______________________________________________________
Error: (/home/micket/projects/oofem/src/oofemlib/intarray.C:104)
In checkBounds:
array error on index : 7 > 6
_______________________________________________________
stack trace:
  /home/micket/projects/build/debug/liboofem.so : oofem::IntArray::checkBounds(int) const+0x102
  /home/micket/projects/build/debug/liboofem.so : oofem::IntArray::at(int)+0x20
  /home/micket/projects/build/debug/liboofem.so : oofem::VTKXMLExportModule::makeFullTensorForm(oofem::FloatArray&, oofem::FloatArray const&)+0x53
  /home/micket/projects/build/debug/liboofem.so : oofem::VTKXMLExportModule::getNodalVariableFromIS(oofem::FloatArray&, oofem::Node*, oofem::TimeStep*, oofem::InternalStateType, int)+0x3f5
  /home/micket/projects/build/debug/liboofem.so : oofem::VTKXMLExportModule::exportIntVars(oofem::VTKPiece&, oofem::IntArray&, oofem::IntArray&, int, oofem::TimeStep*)+0x14b
  /home/micket/projects/build/debug/liboofem.so : oofem::VTKXMLExportModule::setupVTKPiece(oofem::VTKPiece&, oofem::TimeStep*, int)+0x3f1
  /home/micket/projects/build/debug/liboofem.so : oofem::VTKXMLExportModule::doOutput(oofem::TimeStep*, bool)+0x1ed
  /home/micket/projects/build/debug/liboofem.so : oofem::ExportModuleManager::doOutput(oofem::TimeStep*, bool)+0xbe
  /home/micket/projects/build/debug/liboofem.so : oofem::EngngModel::doStepOutput(oofem::TimeStep*)+0x60

So that's either the stress or strain tensor that have an unexpected length.

This is probably due to the rotational part not being handled correctly. I will look closer at this and try to fix it.

Re: Crash of a program

Hi, Mikael

I try to solve more simple problem like this and now there is no crash of a program, but null results.

Help me, please

Post's attachments

prob2.in 5.65 kb, 6 downloads since 2015-03-30 

You don't have the permssions to download the attachments of this post.

Re: Crash of a program

Always run with debug mode first. Just because there are no crashes doesn't mean anything in practice. It might just be luck.

In this new file, you do
NodalLoad 2 loadTimeFunction 1 Components 3 0.0 0.0 -2.25
but it should have 6 components.

Re: Crash of a program

6 components of load don't solve this problem. May be I have problem with dofs?

Can you explain me how to run with debug mode, please?

Re: Crash of a program

Compile OOFEM with build type set to "debug" in CMake (this is the default).


This is the same problem as with the plane stress rotation element; They don't properly export stress. I know what needs to be done to fix this, but I haven't had time yet.

Re: Crash of a program

I fixed the bug in trplanrot. It's on github.com/Micket/oofem now, and will be merged in with oofem.org next time i do so.

When it comes to CCTPlate and other shells/plates, you have to ask for IST_ShellForceTensor, IST_ShellMomentumTensor, IST_ShellStrainTensor, IST_ShellCurvatureTensor.

Exporting the stress and strain for a plate doesn't quite make sense (though, we could take the midplane stress/strain. I'm not sure what would be best there, so I haven't modified it yet.

Re: Crash of a program

Hi, Mikael

There is some problems with input file. When I run it, I get error:

Skyline :: buildInternalStructure - Can't allocate: 406897234.

I can't load file, it's too large.

DSlab.out
Concrete slab -> pure compression. Input mesh is DSlab.unv
LinearStatic nsteps 1 nmodules 1
vtkxml tstep_all domain_all primvars 1 1 vars 3 9 10 11 stype 0
domain 3dshell
OutputManager tstep_all dofman_all element_all
ndofman 11362 nelem 22168 ncrosssect 1 nmat 1 nbc 2 nic 0 nltf 1
node 1     coords 3 3.43     19.275   0        bc 6 1 1 1 1 1 1
node 2     coords 3 5.3      19.275   0        bc 6 1 1 1 1 1 1
node 3     coords 3 5.3      21.075   0        bc 6 1 1 1 1 1 1
node 4     coords 3 2.8      21.075   0        bc 6 1 1 1 1 1 1
node 5     coords 3 2.8      19.275   0        bc 6 1 1 1 1 1 1
node 6     coords 3 0        19.275   0        bc 6 1 1 1 1 1 1
node 7     coords 3 0        24.9     0        bc 6 1 1 0 0 0 1
node 8     coords 3 -0.36    24.9     0        bc 6 1 1 0 0 0 1
node 9     coords 3 -0.36    33.2     0        bc 6 1 1 0 0 0 1
node 10    coords 3 0        33.2     0        bc 6 1 1 0 0 0 1
...
node 11353 coords 3 2.0088   0.32759  0        bc 6 1 1 0 0 0 1
node 11354 coords 3 11.5339  0.303946 0        bc 6 1 1 0 0 0 1
node 11355 coords 3 15.8346  0.636013 0        bc 6 1 1 0 0 0 1
node 11356 coords 3 15.8357  2.53848  0        bc 6 1 1 0 0 0 1
node 11357 coords 3 5.23015  19.122   0        bc 6 1 1 0 0 0 1
node 11358 coords 3 11.9164  0.291047 0        bc 6 1 1 0 0 0 1
node 11359 coords 3 1.81675  0.291929 0        bc 6 1 1 0 0 0 1
node 11360 coords 3 1.81249  0.136313 0        bc 6 1 1 0 0 0 1
node 11361 coords 3 11.8862  0.143924 0        bc 6 1 1 0 0 0 1
node 11362 coords 3 11.5658  0.144777 0        bc 6 1 1 0 0 0 1
CCTplate3D 555   nodes 3   3252 137 138 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 556   nodes 3   136 137 5532 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 557   nodes 3   3252 138 1224 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 558   nodes 3   3254 1224 3253 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 559   nodes 3   594 3254 3255 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 560   nodes 3   3255 140 3281 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 561   nodes 3   3253 140 3255 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 562   nodes 3   3253 3255 3254 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 563   nodes 3   1224 139 3253 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 564   nodes 3   3253 139 140 mat 1 crosssect 1 BodyLoads 1 2
...
CCTplate3D 22714 nodes 3   11343 11337 11349 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 22715 nodes 3   11338 3250 11350 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 22716 nodes 3   11335 11350 3249 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 22717 nodes 3   1223 11338 11335 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 22718 nodes 3   11350 11335 11338 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 22719 nodes 3   11349 3249 11350 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 22720 nodes 3   11348 11350 3250 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 22721 nodes 3   3251 11349 11348 mat 1 crosssect 1 BodyLoads 1 2
CCTplate3D 22722 nodes 3   11350 11348 11349 mat 1 crosssect 1 BodyLoads 1 2
SimpleCS 1 thick 0.22
IsoLE 1 d 2350. E 30.e6 n 0.2 talpha 1.0
BoundaryCondition 1 loadTimeFunction 1 prescribedvalue 0.0
DeadWeight 2 loadTimeFunction 1 Components 6 0.0 0.0 -0.23 0.0 0.0 0.0 valtype 2
ConstantFunction 1 f(t) 1.0

Re: Crash of a program

For large systems, don't use the skyline matrix. The skyline matrix is highly dependent on the nodal numbering. This can be improved by numbering the nodes differently, or using the automatic renumbering, but in many cases, the best numbering is still not good enough to use with a skyline matrix sparse structure.

You have ~68k DOFs in this system, and with this numbering you got a fill rate of ~18% in the matrix. This matrix type don't do for this system. You need PETSc or DSS