Hi all
I have 10,000 water molecule (TIP3P) in my system, I used kspace style with the following interactions:
pair_style lj/cut/coul/long 8 10
kspace_style pppm 1e-4
to run 1,000,000 time step I should use about 3 days. Do you know any suggestion to run faster?
Thanks alot for your time and attention
Best Regards
Elena
Hi all
I have 10,000 water molecule (TIP3P) in my system, I used kspace style with
the following interactions:pair_style lj/cut/coul/long 8 10
kspace_style pppm 1e-4
are you sure about the 8 angstrom LJ cutoff?
that is mightly short?
to run 1,000,000 time step I should use about 3 days. Do you know any
suggestion to run faster?
use more processors?
you may be able to tweak the performance by adjusting
the coulomb cutoff (with long range electrostatics you
get to choose how much work is done in real space and
reciprocal space and there is an optimal cutoff value)
3 days are not _that_ bad. there are projects i know of
that need to be running for accumulated run times of
weeks. 3 days is a long weekend...
axel.