Error with Voronoi occupation in version 30Jul16

LAMMPS LAMMPS Mailing List Mirror
1

Dear LAMMPS users,

I just installed last stable version of LAMMPS from 30July2016. I compiled it with GPU and Voronoi packages. Compilation went smooth.
I tried an example that works with version 14May2016.

It is a simple example where I define a box of Fe atoms with ppp or ppf boundary conditions. After creation of the box and of atoms I proceed to the first invocation of Voronoi tessalation as follows:

INITIAL VORONOI TESSELATION

compute v1 all voronoi/atom occupation
compute r0 all reduce sum c_v1[1]
compute r1 all reduce sum c_v1[2]

thermo_style custom c_r0 c_r1
run 0

This works with version 14May2016. With version 30Jul2016, the simulation stops at this stage with the following error message:

r0 r1
ERROR on proc 0: Failed to allocate -4 bytes for array voronoi/atom:occvec (…/memory.cpp:56)
ERROR on proc 3: Failed to allocate -4 bytes for array voronoi/atom:occvec (…/memory.cpp:56)
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 3
Cuda driver error 4 in call at file ‘geryon/nvd_device.h’ in line 124.
application called MPI_Abort(MPI_COMM_WORLD, -1) - process 3
ERROR on proc 2: Failed to allocate -4 bytes for array voronoi/atom:occvec (…/memory.cpp:56)
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 2
ERROR on proc 1: Failed to allocate -4 bytes for array voronoi/atom:occvec (…/memory.cpp:56)
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 1
Cuda driver error 4 in call at file ‘geryon/nvd_device.h’ in line 124.
application called MPI_Abort(MPI_COMM_WORLD, -1) - process 2

Christophe

2

Dear LAMMPS users,

I just installed last stable version of LAMMPS from 30July2016. I compiled
it with GPU and Voronoi packages. Compilation went smooth.
I tried an example that works with version 14May2016.

It is a simple example where I define a box of Fe atoms with ppp or ppf
boundary conditions. After creation of the box and of atoms I proceed to the
first invocation of Voronoi tessalation as follows:

# INITIAL VORONOI TESSELATION

compute v1 all voronoi/atom occupation
compute r0 all reduce sum c_v1[1]
compute r1 all reduce sum c_v1[2]

thermo_style custom c_r0 c_r1
run 0

This works with version 14May2016. With version 30Jul2016, the simulation
stops at this stage with the following error message:
r0 r1
ERROR on proc 0: Failed to allocate -4 bytes for array voronoi/atom:occvec
(../memory.cpp:56)

computing occupation requires an atom map now. please insert:

atom_modify map hash

and try again.

axel.

3

Dear LAMMPS users,

I just installed last stable version of LAMMPS from 30July2016. I compiled
it with GPU and Voronoi packages. Compilation went smooth.
I tried an example that works with version 14May2016.

It is a simple example where I define a box of Fe atoms with ppp or ppf
boundary conditions. After creation of the box and of atoms I proceed to the
first invocation of Voronoi tessalation as follows:

INITIAL VORONOI TESSELATION

compute v1 all voronoi/atom occupation
compute r0 all reduce sum c_v1[1]
compute r1 all reduce sum c_v1[2]

thermo_style custom c_r0 c_r1
run 0

This works with version 14May2016. With version 30Jul2016, the simulation
stops at this stage with the following error message:
r0 r1
ERROR on proc 0: Failed to allocate -4 bytes for array voronoi/atom:occvec
(…/memory.cpp:56)

computing occupation requires an atom map now. please insert:

atom_modify map hash

It works ! Thanks !

Christophe

4

There should be an error message in compute voronoi/atom for

not using a map in this case - Axel, want to add it?

Steve

5

There should be an error message in compute voronoi/atom for
not using a map in this case - Axel, want to add it?

already done.

https://github.com/akohlmey/lammps/commit/16323ba391eaa8cb07559744ea3e9cb2f70ed726

axel.