segmentation fault; dbg

Dear Dr Axel,

Thank You very much for Your previous comments on segmentation fault debugging;
I am now my working with a new configuration, and again I got a segmentation fault and used the commands You recommended, with the following output (dbg):

#0 0xb7d47c33 in ____strtod_l_internal (nptr=0x0, endptr=0x0, group=0,
     loc=0xb7e6c3a0) at strtod_l.c:530
#1 0xb7d45510 in strtod (nptr=0x0, endptr=0x0) at strtod.c:70
#2 0x08053ca6 in LAMMPS_NS::atom::data_atoms(int, char*) ()
#3 0x081e0c11 in LAMMPS_NS::ReadData::atoms() ()
#4 0x081e1981 in LAMMPS_NS::ReadData::command(int, char**) ()
#5 0x08149133 in LAMMPS_NS::Input::execute_command() ()
#6 0x081496f3 in LAMMPS_NS::Input::file() ()
#7 0x08151aac in main ()

this is how lammps runs into this:

   1 = max bonds/atom
   1 = max angles/atom
Reading data file ...
   orthogonal box = (-200 -200 -200) to (200 200 200)
   1 by 1 by 1 processor grid
Segmentation fault

I assume that bonds and angles sections do not have errors but something else does this? Can it be due to the fact that atoms are too close to each other (some are closer than the cut-off that I am using.)

Sincerely Yours,
Anna

Dear Dr Axel,

Thank You very much for Your previous comments on segmentation fault
debugging;
I am now my working with a new configuration, and again I got a
segmentation fault and used the commands You recommended, with the
following output (dbg):

#0 0xb7d47c33 in ____strtod_l_internal (nptr=0x0, endptr=0x0, group=0,
loc=0xb7e6c3a0) at strtod_l.c:530
#1 0xb7d45510 in strtod (nptr=0x0, endptr=0x0) at strtod.c:70
#2 0x08053ca6 in LAMMPS_NS::atom::data_atoms(int, char*) ()
#3 0x081e0c11 in LAMMPS_NS::ReadData::atoms() ()
#4 0x081e1981 in LAMMPS_NS::ReadData::command(int, char**) ()
#5 0x08149133 in LAMMPS_NS::Input::execute_command() ()
#6 0x081496f3 in LAMMPS_NS::Input::file() ()
#7 0x08151aac in main ()

this is how lammps runs into this:

1 = max bonds/atom
1 = max angles/atom
Reading data file ...
orthogonal box = (-200 -200 -200) to (200 200 200)
1 by 1 by 1 processor grid
Segmentation fault

I assume that bonds and angles sections do not have errors but something
else does this? Can it be due to the fact that atoms are too close to
each other (some are closer than the cut-off that I am using.)

no. this looks as if the code has problems reading your atom data.
since your executable seems to be compiled without debug info,
it does not print the exact line number where it crashed. also you
should check the file. perhaps there is a number/ line too few.
but there is a number of other possibilities, including a memory
management bug in a very different part of the code.

this is difficult to track down any further without seeing the exact input
and data file.

axel.

Dear Dr Axel,

I have checked the data file, and indeed I found some "blank" lines, however, when I removed them, I still got the segmentation fault. I am attaching the files to this e-mail; You can have a look at it if You have time. Thank You very much in advance!

SIncerely Yours,
Anna

initial_configuration2.tar.gz (527 KB)

rosette.tar.gz (549 Bytes)

dear anna

Dear Dr Axel,

I have checked the data file, and indeed I found some "blank" lines,
however, when I removed them, I still got the segmentation fault. I am

check out line 29132. it is missing the coordinates.

29108 1 1 4.1109425 4.45962376 145.535
29109 1 1 4.53158825 4.5527051 145.54
29110 1 1
29111 1 1 4.44776986 4.52739413 145.545
29112 1 1 3.88431544 4.34045678 145.55
29113 1 1 2.93535828 3.97509299 145.555

cheers,
    axel.

Dear Dr Axel,

Thank You so very much for pointing that out! I have now corrected the mistake! I have tried to run it again, and the segmentation fault is gone, but the simulation is still flying apart ( ERROR on proc 0: Bond atoms 133 134 missing on proc 0 at step 1) after dump.0 has been created, and I am now thinking whether I can do something to that or whether I should change my initial configuration.

Thank You so very much again!
Sincerely Yours,
-Anna

Dear Dr Axel,

Thank You so very much for pointing that out! I have now corrected the
mistake! I have tried to run it again, and the segmentation fault is gone,
but the simulation is still flying apart ( ERROR on proc 0: Bond atoms 133
134 missing on proc 0 at step 1) after dump.0 has been created, and I am now
thinking whether I can do something to that or whether I should change my
initial configuration.

your initial configuration is for very high potential energy.
you can try to first run a minimization to remove as much as possible of that
before starting a regular MD.

axel.

Dear Dr Axel,

Thank You very much! I am studying the possible minimisation options that LAMMPS offers!

Sincerely Yours,
Anna