Ercoftac centrifugal pump case study bug, Openfoam 1.6-ext (Floating point exception)
 

Hey,

I am trying to run the ERCOFTAC centrifugal pump OpenFOAM case-study shared on the openFoamwiki by people from Hydro Quebec and chalmers university of Technology.
I am using OpenFoam-1.6-EXT and the solver MRFSimpleFoam.
In the first iterations, the compilation is going well (I want to run the case on 5000 iterations) but later (at about 1400 iterations), there is a bug and the compilation is stopped.

Here is what is written on the log file:

smoothSolver: Solving for Ux, Initial residual = 1.4911964e-10, Final residual = 1.4911964e-10, No Iterations 0
smoothSolver: Solving for Uy, Initial residual = 8.9358568e-10, Final residual = 8.9358568e-10, No Iterations 0
GAMG: Solving for p, Initial residual = 0.98305543, Final residual = 0.0170324, No Iterations 1
time step continuity errors : sum local = 2.9330002e+41, global = 1.7750351e+29, cumulative = 1.7750353e+29
smoothSolver: Solving for epsilon, Initial residual = 6.4743276e-16, Final residual = 6.4743276e-16, No Iterations 0
smoothSolver: Solving for k, Initial residual = 6.2646515e-12, Final residual = 6.2646515e-12, No Iterations 0
ExecutionTime = 711.17 s ClockTime = 1455 s

GGI pair (GGI_INT, GGI_EXT) : 8.2404281e+25 2.921623e+26 Diff = 1.8559864e+25 or 22.522937 %
Time = 1413

smoothSolver: Solving for Ux, Initial residual = 2.7363363e-22, Final residual = 2.7363363e-22, No Iterations 0
smoothSolver: Solving for Uy, Initial residual = 4.2314383e-22, Final residual = 4.2314383e-22, No Iterations 0
GAMG: Solving for p, Initial residual = 4.3439669e-06, Final residual = 4.1146215e-08, No Iterations 1
time step continuity errors : sum local = 3.6948953e+68, global = -5.2860233e+65, cumulative = -5.2860233e+65
smoothSolver: Solving for epsilon, Initial residual = 1, Final residual = 0.05612422, No Iterations 4
[sycomore:01822] *** Process received signal ***
[sycomore:01822] Signal: Floating point exception (8)
[sycomore:01822] Signal code: (-6)
[sycomore:01822] Failing at address: 0x4cd0000071e
[sycomore:01822] [ 0] /lib64/libc.so.6 [0x3e19a30280]
[sycomore:01822] [ 1] /lib64/libc.so.6(gsignal+0x35) [0x3e19a30215]
[sycomore:01822] [ 2] /lib64/libc.so.6 [0x3e19a30280]
[sycomore:01822] [ 3] /usr1/appli/openfoam/OpenFOAM-1.6-EXT/OpenFOAM-1.6-ext/lib/linux64GccDPOpt/libOpenFOAM.so(_ZN4Foam19GaussSeidelSmoother6smoot hERNS_5FieldIdEERKNS_9lduMatrixERKS2_RKNS_10FieldF ieldIS1_dEERKNS_8UPtrListIKNS_17lduInterfaceFieldE EEhi+0x4b0) [0x2b67f41e0b90]
[sycomore:01822] [ 4] /usr1/appli/openfoam/OpenFOAM-1.6-EXT/OpenFOAM-1.6-ext/lib/linux64GccDPOpt/libOpenFOAM.so(_ZNK4Foam19GaussSeidelSmoother6smoo thERNS_5FieldIdEERKS2_hi+0x28) [0x2b67f41e0d38]
[sycomore:01822] [ 5] /usr1/appli/openfoam/OpenFOAM-1.6-EXT/OpenFOAM-1.6-ext/lib/linux64GccDPOpt/libOpenFOAM.so(_ZNK4Foam12smoothSolver5solveERNS_5 FieldIdEERKS2_h+0x2b0) [0x2b67f41da980]
[sycomore:01822] [ 6] /usr1/appli/openfoam/OpenFOAM-1.6-EXT/OpenFOAM-1.6-ext/lib/linux64GccDPOpt/libfiniteVolume.so(_ZN4Foam8fvMatrixIdE5solveERKNS _10dictionaryE+0x16b) [0x2b67f337158b]
[sycomore:01822] [ 7] /usr1/appli/openfoam/OpenFOAM-1.6-EXT/OpenFOAM-1.6-ext/lib/linux64GccDPOpt/libincompressibleRASModels.so(_ZN4Foam5solveIdEENS _9lduMatrix17solverPerformanceERKNS_3tmpINS_8fvMat rixIT_EEEE+0x52) [0x2b67f26a9e52]
[sycomore:01822] [ 8] /usr1/appli/openfoam/OpenFOAM-1.6-EXT/OpenFOAM-1.6-ext/lib/linux64GccDPOpt/libincompressibleRASModels.so(_ZN4Foam14incompress ible9RASModels8kEpsilon7correctEv+0x10fb) [0x2b67f269f6ab]
[sycomore:01822] [ 9] MRFSimpleFoam [0x41533a]
[sycomore:01822] [10] /lib64/libc.so.6(__libc_start_main+0xf4) [0x3e19a1d974]
[sycomore:01822] [11] MRFSimpleFoam(_ZNK4Foam11regIOobject11writeObjectE NS_8IOstream12streamFormatENS1_13versionNumberENS1 _15compressionTypeE+0xc1) [0x412859]
[sycomore:01822] *** End of error message ***
--------------------------------------------------------------------------
mpirun noticed that process rank 0 with PID 1822 on node sycomore.localdomain exited on signal 8 (Floating point exception).


I have noticed that before the bug, the values of the amount of flux going through the GGI patched pair GGI_INT and GGI_ext are completly different. May be there is a problem with the mesh and GGI...

Does anybody have the same problem?
The ercofact case study has been run with OpenFoam-1.5-dev, so do you think i have a problem with the version of openFoam, as i am using the 1.6-ext?

I need your help!!

Ariane

Hi,

did you get the case running in OF16ext or did you switched to OF15dev?

Has anyone here tried to run the case in OF16ext?

Best regards,
Hannes