From NWChem
Viewed 3803 times, With a total of 2 Posts
|
|
|
|
| 12:41:18 PM PDT - Mon, Mar 14th 2011 |
|
It's not clear from the documentation what scf algorithm is implemented for DFT calculation. Is it direct diagonalization of Fock matrix with diis convergence acceleration or direct minimization of the energy functional with the diis method? In case of convergence failure, is there another more robust method like preconditioned conjugate gradient as a last resort?
Any help is appreciated.
|
|
|
-
Niri Forum:Admin, Forum:Mod, NWChemDeveloper, bureaucrat, sysop
|
|
|
Gets Around
Threads 4
Posts 116
|
|
| 8:06:58 AM PDT - Wed, Mar 16th 2011 |
|
Hi,
Yes. The DFT uses diagonalization with DIIS convergence. We have an implementation which uses a CG scheme but this is part of the development. I plan to release it soon. In the meanwhile you can use the CG within our HF module and then use these vectors as a starting guess for the DFT stage. Check the vectors input keyword.
Here's a snippet of an example:
scf
maxiter 100
vectors output hf.movecs
end
task scf
dft
iterations 100
xc pbe0
vectors input hf.movecs output pbe0.movecs
end
task dft
Hope this helps. Let me know if you need further clarification.
-Niri
|
|
|
AWC's:
2.5.10 MediaWiki - Stand Alone Forum Extension
Forum theme style by: AWC
