Aman,
Your “-pk kokkos device 1” command isn’t quite right and that is why you are getting the original error. Please see the docs here http://lammps.sandia.gov/doc/package.html. I would recommend using an unmodified input script (no /kk commands in the input script) along with:
lmp_kokkos_cuda_mpich -pk kokkos newton on neigh half comm no -k on g 1 -sf kk < in.CHO
Often the half neighbor list is better than the default full even on the GPU. There is also a new “-pk kokkos neigh/qeq” option that you can try out as well, sometimes it is better to use a full neighbor list with fix qeq/reax/kk and a half neighbor list for pair reax/c/kk. You will need to update your version of LAMMPS though.
Stan
Thanks for your suggestions,
I installed the latest LAMMPS (10th March, 2017) version and tried the same input script given in the example without any modification. I used the command suggested by Stan but I got the same error:
Neighbor list info ...
update every 10 steps, delay 0 steps, check no
max neighbors/atom: 2000, page size: 100000
master list distance cutoff = 12
ghost atom cutoff = 12
binsize = 6, bins = 5 5 5
2 neighbor lists, perpetual/occasional/extra = 2 0 0
(1) pair reax/c/kk, perpetual
attributes: half, newton off, ghost, kokkos_device
pair build: half/bin/ghost/kk/device
stencil: half/ghost/bin/3d/newtoff
bin: kk/device
(2) fix qeq/reax, perpetual, copy from (1)
attributes: half, newton off, ghost, kokkos_device
pair build: copy/kk/device
stencil: none
bin: none
Setting up Verlet run ...
Unit style : real
Current step : 0
Time step : 0.25
terminate called after throwing an instance of 'std::runtime_error'
what(): cudaMalloc( &ptr, arg_alloc_size ) error( cudaErrorMemoryAllocation): out of memory ../../lib/kokkos/core/src/Cuda/Kokkos_CudaSpace.cpp:176
Traceback functionality not available
Aborted (core dumped)
Please give some advice.
Thanks,
Aman