Hello
I am conducting a simulation with 4000 particles using the SNAP Ni Potential to calculate the enthalpy and density of the system. The system is being heated from 300K to 2700K, and I am studying the phase changes. The first time I ran the simulation, I used the Ni_Zuo_JPCA2020 with Kokkos and it took about 10 hours to finish. However, when I ran the simulation using OpenKIM with the same potential, it has been running for more than 3 days and is less than halfway done. How can I speed up the simulation?
This is the Script
variable nrep equal 10
variable a equal 3.527
kim init SNAP_ZuoChenLi_2019_Ni__MO_365106510449_000 metal
variable nx equal ${nrep}
variable ny equal ${nrep}
variable nz equal ${nrep}
boundary p p p
lattice fcc $a
region box block 0 ${nx} 0 ${ny} 0 ${nz}
create_box 1 box
create_atoms 1 box
mass 1 58.69
kim interactions Ni
thermo 100
thermo_style custom step density temp press enthalpy
timestep 1.0e-3
neighbor 1.0 bin
neigh_modify once no every 1 delay 0 check yes
velocity all create 300.0 1523945 dist gaussian
variable i loop 50
label loopa
variable tx equal 300+50*($i-1)
log heating_ni_t${tx}.out
fix 1 all npt temp ${tx} ${tx} 0.01 iso 0.0 0.0 0.1
run 40000
unfix 1
next i
jump in.heating loopa
variable i loop 50
label loopb
variable tx equal ${tx}-50
log cooling_ni_t${tx}.out
fix 1 all npt temp ${tx} ${tx} 0.01 iso 0.0 0.0 0.1
run 40000
unfix 1
next i
jump in.heating loopb
The command i use to run the simulation is
nohup nice mpirun -n 1 lmp -k on g 1 -sf kk -pk kokkos cuda/aware on neigh half newton on -in in.heating >./out.heating &
Thank you and greetings!