| 10:54:30 PM PDT - Thu, May 21st 2015 |
|
Hi,
I'm running a nwchem code to get the dft energy between 2 monomers with BSSE correction.
The code was initially taking many errors like "failed to converge". But when I changed grid input from 'coarse' to 'xfine', it did converge. Although taking a way more time on 32 processors.
So, what I want to know is that:
What is the reason that the convergence is getting failed on "coarse" but not in "xfine" ? Am I taking the right decision on it ?
Will the calculation speed increase if I just improve the number of cpus ?
A little idea on these points will be highly appreciated.
thank you.
|
|
|