Dear Lammps users,
I am using hybrid paristyle( airbeo + eam + Lj/cut) to simulate Aluminum- CNT interaction. In this case my simulations run in Serial. If I use omp package, It runs in parallel but the results are wrong. How can I solve this problem?
Thanks,
Samaneh
Dear Lammps users,
I am using hybrid paristyle( airbeo + eam + Lj/cut) to simulate Aluminum-
CNT interaction. In this case my simulations run in Serial. If I use omp
package, It runs in parallel but the results are wrong. How can I solve this
problem?
which version of LAMMPS is this with? and on what platform?
what do you mean by "wrong"?
axel.
Dear Axel,
I am using LAMMPS (30 Nov 2016) on my fedora25.
Cohesive energy of Carbon is about -7.3 eV/atom but when I run with omp it is -2.1 eV/atom. I call it " wrong" result.
Bests,
Samaneh
Dear Axel,
I am using LAMMPS (30 Nov 2016) on my fedora25.
Cohesive energy of Carbon is about -7.3 eV/atom but when I run with omp it is -2.1 eV/atom. I call it " wrong" result.
it is still not clear what you are comparing here. how are you
computing this and where is this -7.3 eV coming from?
axel.
Sorry if I am not clear enough.
I am comparing the energy of carbon when I run an identical input script with and without omp package.
without omp package energy of carbon is -7.3 eV which is expected comparing with literature. But with omp package energy of carbon is -2.1 eV.
Sorry if I am not clear enough.
I am comparing the energy of carbon when I run an identical input script with and without omp package.
without omp package energy of carbon is -7.3 eV which is expected comparing with literature. But with omp package energy of carbon is -2.1 eV.
ok. i don't see a different behavior, so in order to track this down,
i need to have a complete input deck that reproduces it. thus can
please provide me with your complete input, makefile settings, and the
exact command line you were using for each run.
thanks,
axel.