Gromacs vs lammps

LAMMPS LAMMPS Mailing List Mirror
1

Dear all:

I am a regular gromacs user, recently, i need to run very large system approx 2million atoms. I am considering to shift to lammps if i can gain some speed up. Do anybody have any benchmark report comparing gromacs vs lammps for normal cluster(without any accelerator).

–Masrul

2

Dear all:

I am a regular gromacs user, recently, i need to run very large system
approx 2million atoms. I am considering to shift to lammps if i can gain
some speed up.

what do you mean by "gain some speed up"?

Do anybody have any benchmark report comparing gromacs vs
lammps for normal cluster(without any accelerator).

for what kind of system? and at what system size for how many CPU cores?

since you know gromacs yourself, your best bet is to take one of the
benchmark examples for LAMMPS, translate it to gromacs and compare
what you find with what is available here (or run LAMMPS on your
cluster with those same inputs, too):
http://lammps.sandia.gov/bench.html

axel.

3

In addition to Axel's advice on simply benchmarking the problem - from
my own experience w/Gromacs and LAMMPS: you say "approx 2 million
atoms". Does this mean 2 million "significant atoms" or 1.95 million
"rigid water" + 0.05 million "significant atoms"?

Due to the special optimization for rigid water in Gromacs, this
questions is, imho, to be answered first.

Regards

M.