Hi Azar,

The relaxation information (energies/forces) may or may not be sensitive to k 
grid depending on your system.

If you are not sure, try doing the same calculation on different k grid sizes 
and look how the output changes.

--------------------------------------------------------
Aaron Thong
PhD research student
Department of Materials
B321 Desk 20
Imperial College London
--------------------------------------------------------


On 25 Nov 2015, at 16:17, azar ostovan wrote:

Dear Aaron Thong
Thank you so much
I try such kgrid just for siesta calculation after structural relaxation
in transport calculation the z-direction will be extended to 100.
So if I got correctly you mean the siesta relaxation information are not 
sensitive to the number of kgrid?
Cheers
azar

On Wed, Nov 25, 2015 at 7:34 PM, Thong, Aaron 
<[email protected]<mailto:[email protected]>> wrote:
Hi Azar,

Would be helpful to think about what you are trying to calculate. The MP grid 
can differ across calculations (i.e. SIESTA will not complain. Advanced users 
please correct me if I'm wrong here). But this may not mean that the answer you 
obtain is correct/physical.

Using 1 k point would save time but also you lose out information on electron 
densities/transmission/properties (depending on what job you are running) at 
other K points. This may or may not be important depending on your system.

--------------------------------------------------------
Aaron Thong
PhD research student
Department of Materials
B321 Desk 20
Imperial College London
--------------------------------------------------------


On 25 Nov 2015, at 15:46, azar ostovan wrote:

Dear transiesta users
.
Do the number of kgrid_Monkhorst_Pack block ( or periodicity in each direction) 
could differ in structural relaxation, transiesta and tbtrans calculations? Or 
it must be the same?
I mean can I for time saving used the below block in the optimization and 
transiesta
%block kgrid_Monkhorst_Pack
1   0   0   0.0
0   1   0   0.0
0   0   1  0.0
%endblock kgrid_Monkhorst_Pack
but for tbtrans extend the structure in z-direction?
Warm regards
azar




Responder a