This thread is locked.Only browsing is available.
Top Page > Browsing
STM Simulation doesn't work when solver is Cluster
Date: 2020/12/27 21:04
Name: Kelvin

Dear OpenMX,

I notice that the STM simulation as exemplified in cannot work for Cluster solver (which we are forced to use when we use Kgrid 1 1 1 (i.e. only gamma point sampling)). The calculation outputs pden.cube that is zero on all voxels.

Could you suggest some workaround (e.g. copy-pasting some part in Band_DFT_dosout.c to Cluster_DFT_dosout.c)?

How I found it:
I ran the STM simulation example as instructed in and I can reproduce the results displayed there. When I changed the Kgrid to 1 1 1, the solver was automatically changed to Cluster and the pden.cube was zero on all voxels.

Hoping to seek your advice on this; because when we simulate the STM image for molecules on surface, it is often not possible to evaluate beyond gamma-point. Thanks so much for your help.

Page: [1]

Re: STM Simulation doesn't work when solver is Cluster ( No.1 )
Date: 2020/12/28 02:42
Name: Naoya Yamaguchi


Please try the following and do "make" again to disable the automatic change.
$ sed -ie 's/Solver = 2/\/\/Solver = 2/' Input_std.c

Naoya Yamaguchi

Page: [1]