Seg error trying to combine Gay-Berne and dipoles

[Margaret Rosenberg]

Dear all,

I'm attempting to reproduce a simple Gay-Berne ferrofluid and getting 
some odd segmentation errors. I've attached an attempt at a minimal 
working example* below: any advice would be appreciated. The error 
message reads as follows:

[cpk01:64078] *** Process received signal ***
[cpk01:64078] Signal: Segmentation fault: 11 (11)
[cpk01:64078] Signal code: Address not mapped (1)
[cpk01:64078] Failing at address: 0x7f85d9bd5000
[cpk01:64078] [ 0] 0   libsystem_platform.dylib            
0x00007fff718535fd _sigtramp + 29
[cpk01:64078] [ 1] 0   ???                                 
0x00007f89d7c30220 0x0 + 140230007128608
[cpk01:64078] [ 2] 0   EspressoCore.dylib                  
0x0000000109d691b6 _Z18dp3m_dipole_assignRK13ParticleRange + 326
[cpk01:64078] [ 3] 0   EspressoCore.dylib                  
0x0000000109d74bd7 _ZN6Dipole21calc_long_range_forceERK13ParticleRange + 
359
[cpk01:64078] [ 4] 0   EspressoCore.dylib                  
0x0000000109cba83c _Z10force_calcR13CellStructure + 1052
[cpk01:64078] [ 5] 0   EspressoCore.dylib                  
0x0000000109ccf1ce _Z12integrate_vvii + 702
[cpk01:64078] [ 6] 0   EspressoCore.dylib                  
0x0000000109c956d9 _ZL19mpi_integrate_slaveii + 9
[cpk01:64078] [ 7] 0   EspressoCore.dylib                  
0x0000000109c97b78 
_ZNK13Communication12MpiCallbacks4callINSt3__14plusIiEEiJiiEEENS2_16remove_referenceIDTclfp0_clsr3stdE7declvalIT0_EEclsr3stdE7declvalIS6_EEEEE4typeENS_6Result9ReductionET_PFS6_DpT1_ESE_ 
+ 88
[cpk01:64078] [ 8] 0   EspressoCore.dylib                  
0x0000000109ccf51a _Z16python_integrateibb + 506
[cpk01:64078] [ 9] 0   integrate.so                        
0x000000010c98167e 
_ZL50__pyx_pw_10espressomd_9integrate_10Integrator_9runP7_objectS0_S0_ + 
606
[cpk01:64078] [10] 0   script_interface.so                 
0x000000010a73a71c _ZL29__Pyx_CyFunction_CallAsMethodP7_objectS0_S0_ + 
92
[cpk01:64078] [11] 0   Python                              
0x0000000109135ad3 _PyObject_MakeTpCall + 266
[cpk01:64078] [12] 0   Python                              
0x0000000109137e04 method_vectorcall + 205
[cpk01:64078] [13] 0   Python                              
0x00000001091ddb23 call_function + 403
[cpk01:64078] [14] 0   Python                              
0x00000001091dacb4 _PyEval_EvalFrameDefault + 27452
[cpk01:64078] [15] 0   Python                              
0x00000001091de66b _PyEval_EvalCode + 1998
[cpk01:64078] [16] 0   Python                              
0x00000001091d409d PyEval_EvalCode + 79
[cpk01:64078] [17] 0   Python                              
0x000000010920f0e5 run_eval_code_obj + 110
[cpk01:64078] [18] 0   Python                              
0x000000010920e4dd run_mod + 103
[cpk01:64078] [19] 0   Python                              
0x000000010920d3a1 PyRun_FileExFlags + 241
[cpk01:64078] [20] 0   Python                              
0x000000010920c991 PyRun_SimpleFileExFlags + 271
[cpk01:64078] [21] 0   Python                              
0x00000001092248b8 Py_RunMain + 1839
[cpk01:64078] [22] 0   Python                              
0x0000000109224bf1 pymain_main + 306
[cpk01:64078] [23] 0   Python                              
0x0000000109224c3f Py_BytesMain + 42
[cpk01:64078] [24] 0   libdyld.dylib                       
0x00007fff7165acc9 start + 1
[cpk01:64078] *** End of error message ***

This is using pypresso 4.1.4 on Mac Catalina. I have considered the 
typical case (system exploding due to high energy), but right before the 
seg error I get energies of (for example):

[('total', 162.0677238013178), ('external_fields', 0.0), ('kinetic', 
162.8133351668658), ('bonded', 0.0), (('non_bonded', 0, 0), 
-0.00713096266954227), ('non_bonded', -0.00713096266954227), ('coulomb', 
0.0), (('dipolar', 0), -0.6532714897406531), (('dipolar', 1), 
-0.08520891313778406), ('dipolar', -0.7384804028784371)]

which seems survivable in terms of dipolar interaction strength? 
However, running the same system without dipolar interactions does not 
cause a segmentation fault (purely thermostat + Gay-Berne).

Best,

Margaret

*The full script included a warm up with purely steric interactions & 
steepest descent to distance particles: it would complete this routine 
and then crash with the same seg error once the integrator was changed 
to vv so I cut the section for time. Same with randomising particle 
orientations, starting on a grid for proper spacing etc. The original 
system is even more dense, with a larger box; the Gay-Berne and other 
system parameters are lifted from an older publication that included 
such simulations.

--
Margaret Rosenberg, MSc

Dipolar Soft Matter Group
Faculty of Physics, University of Vienna
Sensengasse 8/18, 1090, Vienna, Austria

Phone: +43-(0)1-4277-73270
https://dsm.univie.ac.at/people/details/user/rosenbergm69/inum/1151/backpid/96885/

simpler_segerror.py
Description: Text document