Dear LAMMPS Users,
Hello. I have modeled a triclinic simulation box and I would like to apply a shear stress to the box subject to NVT ensemble. I was curious if there is a way to use the fix deform command rather than using the fix npt and setting the barostat parameters to only applied shear.
In my LAMMPS input script, I used the following fix commands:
fix 1 all langevin 1.0 1.0 2.0 12345
fix 2 all nve
fix 3 all deform 1 xy variable v_displace v_rate remap v
I was curious if only these three fix command will perform the shear deformation that I want under NVT condition.
I am using triclinic simulation box since the box will be sheared. Therefore, I cannot use *"*volume x volume y" command in my fix deform command. However, I was curious if fix nve will resolve this issue. Thank you.
Sincerely,
Masato Koizumi
Dear LAMMPS Users,
Hello. I have modeled a triclinic simulation box and I would like to apply
a shear stress to the box subject to NVT ensemble. I was curious if there
is a way to use the fix deform command rather than using the fix npt and
setting the barostat parameters to only applied shear.
In my LAMMPS input script, I used the following fix commands:
fix 1 all langevin 1.0 1.0 2.0 12345
fix 2 all nve
fix 3 all deform 1 xy variable v_displace v_rate remap v
I was curious if only these three fix command will perform the shear
deformation that I want under NVT condition.
the best way to find out is to apply the scientific method, set up a
simple test system and observe.
as i just wrote in another email asking a similar style of question, this
is not the oracle of delphi nor an advise-o-mat that will absolve you from
doing the due diligence yourself.
axel.
Hi,
Sorry about the repost but I would really appreciate any comment on this:
I am trying to compile lammps with Kokkos package for KNL architecture.
My compiler is intel/16.0.3 with intel mpi.
I tried this:
make yes-kokkos
make yes-user-reaxc
make kokkos_phi
I am getting this error:
cc1plus: error: bad value (knl) for -march=switch
cc1plus: error: bad value (knl) for -mtune=switch
I also tried with gcc compiler. In this case it compiled successfully but this case performance was poor.
As an example:
mpirun -np 2 ./lmp_kokkos_phi -k on t 2 -sf kk -in in.batio3
It runs but if i increase the number of threads
mpirun -np 2 ./lmp_kokkos_phi -k on t 4 -sf kk -in in.batio3
It starts running but not doing integration part:
Time step : 0.25
WARNING: Fixes cannot send data in Kokkos communication, switching to classic communication (…/comm_kokkos.cpp:382)
with this warning.
I have this cpu 1.40GHz Xeon Phi 7250 (Knights Landing)
Looks like it supposed to have 68 cores. Simulating system contains only 5k atoms.
Regards.