[francesco_b]

Hi All,

running interFoam for a complex problem, I noticed that suddenly deltaT decreases and from e-07 it goes to e-14, then the computation stop with the following lines:

area3@calcolo:~/OpenFOAM/area3-1.4.1/run/tutorials/interFoam$ mpirun noticed that job rank 0 with PID 16148 on node calcolo exited on signal 15 (Terminated).

[1]+ Exit 143 mpirun -np 2 interFoam . BCplacca -parallel > log_bcplacca
area3@calcolo:~/OpenFOAM/area3-1.4.1/run/tutorials/interFoam$ mpirun -np 2 interFoam . BCplacca -parallel > log_bcplacca &
[1] 16283
area3@calcolo:~/OpenFOAM/area3-1.4.1/run/tutorials/interFoam$ [1] #0 Foam::error::printStack(Foam:stream&) in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
[1] #1 Foam::sigFpe::sigFpeHandler(int) in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
[1] #2 Uninterpreted: [0xffffe420]
[1] #3 Foam::Time::adjustDeltaT() in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
[1] #4 Foam::Time::setDeltaT(double)[0] #0 Foam::error::printStack(Foam:stream&) in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
[0] #1 Foam::sigFpe::sigFpeHandler(int) in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
[0] #2 Uninterpreted: [0xffffe420]
[0] #3 Foam::Time::adjustDeltaT() in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
[0] #4 Foam::Time::setDeltaT(double) in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
[1] #5 in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so"
[0] #5 mainmain in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/applications/bin/linuxGccDPOpt/interFoam"
[1] #6 __libc_start_main in "/lib/tl in s/i686/cmov/libc.so.6"
[1] #7 Foam::regIOobject::readIfModified()"/home/area3/OpenFOAM/OpenFOAM-1.4.1/applicat ions/bin/linuxGccDPOpt/interFoam"
[0] #6 __libc_start_main in "/lib/tls/i686/cmov/libc.so.6"
[0] #7 Foam::regIOobject::readIfModified() in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/applications/bin/linuxGccDPOpt/interFoam"
[calcolo:16287] *** Process received signal ***
[calcolo:16287] Signal: Floating point exception (8)
[calcolo:16287] Signal code: (-6)
[calcolo:16287] Failing at address: 0x3f9f
[calcolo:16287] [ 0] [0xffffe440]
[calcolo:16287] [ 1] /home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so(_ZN4Foam6si gFpe13sigFpeHandlerEi+0x61) [0xb703f491]
[calcolo:16287] [ 2] [0xffffe420]
[calcolo:16287] [ 3] /home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so(_ZN4Foam4Ti me9setDeltaTEd+0x2d) [0xb70c856d]
[calcolo:16287] [ 4] interFoam [0x8062708]
[calcolo:16287] [ 5] /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe0) [0xb6d29050]
[calcolo:16287] [ 6] interFoam(_ZN4Foam11regIOobject14readIfModifiedEv+ 0x1f9) [0x805f4e1]
[calcolo:16287] *** End of error message ***
in "/home/area3/OpenFOAM/OpenFOAM-1.4.1/applications/bin/linuxGccDPOpt/interFoam"
[calcolo:16286] *** Process received signal ***
[calcolo:16286] Signal: Floating point exception (8)
[calcolo:16286] Signal code: (-6)
[calcolo:16286] Failing at address: 0x3f9e
[calcolo:16286] [ 0] [0xffffe440]
[calcolo:16286] [ 1] /home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so(_ZN4Foam6si gFpe13sigFpeHandlerEi+0x61) [0xb7046491]
[calcolo:16286] [ 2] [0xffffe420]
[calcolo:16286] [ 3] /home/area3/OpenFOAM/OpenFOAM-1.4.1/lib/linuxGccDPOpt/libOpenFOAM.so(_ZN4Foam4Ti me9setDeltaTEd+0x2d) [0xb70cf56d]
[calcolo:16286] [ 4] interFoam [0x8062708]
[calcolo:16286] [ 5] /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe0) [0xb6d30050]
[calcolo:16286] [ 6] interFoam(_ZN4Foam11regIOobject14readIfModifiedEv+ 0x1f9) [0x805f4e1]
[calcolo:16286] *** End of error message ***
mpirun noticed that job rank 0 with PID 16286 on node calcolo exited on signal 8 (Floating point exception).

Did it stop because deltaT was tto little? If not, Why did it stop?

Is there something I can do to overcome this problem? Maybe refine the mesh could help?

Thanks for your suggestions

Francesco

[mattijs]

Yes, it stopped because deltaT was too small. More specifically adjustDeltaT() in Time.C (in $FOAM_SRC/OpenFOAM/lnInclude/). There is some logic there which divides by deltaT_ to find out when to write.

Refining the mesh will not likely help. More likely a boundary problem.

[gschaider]

It's been my impression in the past, that the adjustable timestep can do horrible things with interFoam: a small droplet increases its velocity and "pulls" the timestep down.
Don't know the reason for that, but usually with a fixed (small enough) time-step these calculations worked without a problem.