errors: segmentation fault and others on certain jobs only

LAMMPS LAMMPS Mailing List Mirror
1

Dear users

I get errors as soon as I submit some jobs but not with others. The difference between the jobs may only be a change in number of atoms or a change in box dimension.
As far as I can see the input and data files are all correct. The errors give no indication of whats wrong. Whether a job runs or gives errors seems almost random!

See errors below:

FIRST ERROR TYPE:
Reading data file …
orthogonal box = (-1000 -1000 -0.7113) to (1000 1000 425.291)
6 by 6 by 1 processor grid
[r2i1n3:mpi_rank_1][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_2][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_3][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_4][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_5][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_6][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_7][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_8][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_9][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_10][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_11][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n3:mpi_rank_0][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_12][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_13][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_14][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_15][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_16][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_17][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_22][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_18][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_19][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_20][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_21][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n1:mpi_rank_23][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_24][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_25][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_26][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_27][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_28][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_29][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_30][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_31][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_32][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_33][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_34][error_sighandler] Caught error: Segmentation fault
(signal 11)
[r2i1n11:mpi_rank_35][error_sighandler] Caught error: Segmentation fault
(signal 11)
mpiexec_raw: Warning: tasks 0-35 died with signal 11 (Segmentation
fault).

SECOND ERROR TYPE:
ERROR on proc 0: Cannot open dump file
[cli_0]: aborting job:
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0
mpiexec_raw: Warning: task 0 exited with status 1.
mpiexec_raw: Warning: tasks 1-35 died with signal 15 (Terminated).

Thanks in advance,

Daniel

2

Dear users

I get errors as soon as I submit some jobs but not with others. The
difference between the jobs may only be a change in number of atoms or a
change in box dimension.
As far as I can see the input and data files are all correct. The errors

but is the code correct? segmentation fault is a very unspecific error
but usually refers to a bug in the code itself, unless it is something
unexpected (or stupid) where LAMMPS has no safeguards for performance
reasons. the location of the error message makes it likely that it is
related to reading the data file.
as always, to have somebody sort it out, you need to verify that this
happens with the latest version of LAMMPS and provide the smallest
possible input deck that can reproduces the issue.

give no indication of whats wrong. Whether a job runs or gives errors seems
almost random!

See errors below:

FIRST ERROR TYPE:
Reading data file ...
  orthogonal box = (-1000 -1000 -0.7113) to (1000 1000 425.291)
  6 by 6 by 1 processor grid
[r2i1n3:mpi_rank_1][error_sighandler] Caught error: Segmentation fault
(signal 11)

[...]

SECOND ERROR TYPE:
ERROR on proc 0: Cannot open dump file
[cli_0]: aborting job:
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0
mpiexec_raw: Warning: task 0 exited with status 1.
mpiexec_raw: Warning: tasks 1-35 died with signal 15 (Terminated).

"cannot open dump file" is a self-explanatory error. you're obviously
trying to write to a file without suitable permissions to the path, or
without sufficient storage capacity or quota.

axel.