pspw convergence problem


Just Got Here
Hi all,

I'm trying to do a pspw optimisation on my system, but it doesn't converge. When I did this with nosymm, it converged easily. Whenever I turned on the symmetry, it ended until the max. iteration number even I run it for a couple of days on parallel. I read many hints for dft optimisation problems, but I haven't seen any for pspw optimize.

I also wonder how much the simulation box size and cutoff are important when doing a pspw calculation. I'm sorry if this question sounds weird, but I'm totally new to Nwchem.

I'm sending my input. Any advice will be appreciated.

echo
title "optimize with pbe96"

scratch_dir /home/user/nwchemtest//sc
permanent_dir /home/user/nwchemtest/perm

start test_pspw
memory stack 1600 mb heap 800 mb global 1600 mb
geometry units angstrom noautoz print xyz
 O      1.009000   -0.012000   -0.011000
N 2.380000 -0.008000 0.007000
C 2.861000 -0.704000 1.248000
H 2.460000 -0.161000 2.108000
H 3.954000 -0.692000 1.244000
H 2.473000 -1.726000 1.225000
C 2.893000 -0.728000 -1.207000
H 2.513000 -0.202000 -2.087000
H 2.507000 -1.750000 -1.173000
H 3.986000 -0.713000 -1.176000
C 2.872000 1.411000 0.000000
H 2.493000 1.887000 -0.909000
H 3.964000 1.393000 0.016000
H 2.468000 1.904000 0.888000
symmetry c3v
end
nwpw
  simulation_cell
SC 15.0
end
xc pbe96
lmbfgs
cutoff 30.0
end
driver
 clear 
maxiter 100
end

task pspw optimize

Gets Around
The geometry being produced by your input is weird. You might want to comment out the symmetry option.


Forum >> NWChem's corner >> Running NWChem