I think the gradient of energy or forces on atoms are calculated
automatically along with single-point total energy. If you do not need
these information, just ignore them.
Yun-Peng
On 10/04/2011 09:35 PM, [email protected] wrote:
Dear Abraham Hmiel,
thank you very much for reply, but I have tried it and it is somehow
not working ... I still can see calculated gradients in my output file.
I'm sending parts of my input and output files ... hope it
will help you to answer me
Best Regards
Rob
input:
SystemName ala benchmark
SystemLabel UU
NumberOfAtoms 121
NumberOfSpecies 9
MD.TypeOfRun CG
MD.NumCGsteps 0
%block ChemicalSpeciesLabel
1 6 C
2 1 H
3 7 N
4 8 O
5 16 S
6 -6 Cg
7 -1 Hg
8 -7 Ng
9 -8 Og
%endblock ChemicalSpeciesLabel
LatticeConstant 1.00 Ang
%block LatticeVectors
60.00 0.000 0.000
0.000 56.00 0.000
0.000 0.000 52.00
%endblock LatticeVectors
AtomicCoordinatesFormat Ang
%block AtomicCoordinatesAndAtomicSpecies
24.350 25.511 11.899 1
23.694 25.640 10.527 1
24.420 25.843 9.543 4
24.624 24.471 12.075 2
23.651 25.839 12.669 2
25.245 26.132 11.933 2
22.403 25.535 10.514 3
21.544 25.599 9.350 1
20.360 26.557 9.488 1
19.824 26.699 10.639 4
20.954 24.151 9.137 1
20.028 24.048 7.933 1
22.175 23.363 9.006 4
21.973 25.416 11.420 2
22.068 25.899 8.442 2
20.406 23.884 10.040 2
...
25.857 26.569 4.724 7
24.371 28.369 8.376 8
24.360 29.512 9.317 6
24.424 27.435 8.757 7
25.360 29.942 9.379 7
24.050 29.168 10.304 7
23.661 30.269 8.960 7
%endblock AtomicCoordinatesAndAtomicSpecies
MeshCutoff 100. Ry
DM.MixingWeight 0.20
DM.NumberPulay 4
SpinPolarized false
xc.functional GGA
xc.authors PBE
SolutionMethod diagon
MaxSCFIterations 50
MeshSubDivisions 4
output:
siesta: eta*DQ = 0.000000
siesta: Emadel = 0.000000
siesta: Ekinion = 0.000000
siesta: Eharris = -6488.578017
siesta: Etot = -6488.582117
siesta: FreeEng = -6488.582117
siesta: Final energy (eV):
siesta: Kinetic = 4646.640512
siesta: Hartree = 34790.552843
siesta: Ext. field = 0.000000
siesta: Exch.-corr. = -1880.049443
siesta: Ion-electron = -73659.976920
siesta: Ion-ion = 29614.250890
siesta: Ekinion = 0.000000
siesta: Total = -6488.582117
siesta: Atomic forces (eV/Ang):
siesta: 1 -0.658910 0.096442 -0.361758
siesta: 2 4.423454 0.483187 -0.487542
siesta: 3 -0.462609 0.248633 -0.407234
siesta: 4 0.151969 -0.568105 0.148233
siesta: 5 -0.216366 0.102353 0.696090
siesta: 6 0.381259 0.245431 -0.129426
siesta: 7 -3.483028 0.292558 -0.295695
siesta: 8 -1.172083 -1.017509 0.040193
siesta: 9 0.180361 -0.404742 4.428685
siesta: 10 0.856933 -0.313271 -2.190146
siesta: 11 1.750654 0.320628 -0.869240
siesta: 12 0.610253 0.013602 0.626675
siesta: 13 -0.827303 0.988147 0.057831
-------------------------------------------------------
Robert,
It is absolutely possible to do this, the total energy is printed a few
times when the run ends.
To calculate single point energy, use:
MD.TypeOfRun cg
MD.NumCGsteps 0
So that atoms are frozen in place, then once the run is finished either
look
for a "siesta: total Energy" line in the output file, or you can use grep
to
look for the specific line in the file. Try:
fgrep "siesta: FreeEng"
if you want the free energy or
fgrep "siesta: Etot"
for the total energy (in many instances, they will be the same). The grep
command used this way will return two values, the first is the
(unconverged)
total energy from the beginning of the run, and the second is the
converged
energy at the end of the cg/scf loop. Towards the end of the output file,
you can also find lines like these:
siesta: Final energy (eV):
siesta: Band Struct. = -28113.082516
siesta: Kinetic = 62511.903467
siesta: Hartree = 1851209.928718
siesta: Ext. field = 0.000000
siesta: Exch.-corr. = -19076.142766
siesta: Ion-electron =-3768244.293672
siesta: Ion-ion = 1753136.700291
siesta: Ekinion = 0.000000
siesta: Total = -120461.903961
Where the energy is partitioned into all of its components. You have to
carry the calculation through to completion, either by satisfying some
atomic force convergence condition or a maximum of conjugate gradient
steps
to get this information, so you won't see it if you are interrupting the
calculation somehow.
Best of luck,
Abraham Hmiel
Katherine Belz Groves Fellow in Nanoscience
Xue Group, College of Nanoscale Science and Engineering at SUNY Albany
On Tue, Oct 4, 2011 at 7:53 AM,<[email protected]> wrote:
Hello SiestaUsers,
I'm new member of SiestaUsers family. I have one small question,
is it possible to calculate only single point (electronic) energy with
siesta ... . In my output I can see also gradients, but I would like to
calculate only energy.
Thanks in advance.
Best Regards
Robert Sedlak
