Messages

5101

General Questions and Answers / Re: Cannot run script in parallel

« on: May 25, 2009, 21:03 »

My primary suggestion would be to generate the structures first, in serial, on your local machine. Then you read in the geometries when you perform the calculations in parallel. That way you have no randomness in parallel. Generating the geometries anyway will not parallelize, so there is no performance difference.

Otherwise it seems you already have a "parallel" random function, the one you posted. I'm not how it can be improved because you need a seed which is common on all nodes, but you have no way of communicating it between the nodes, so it must be taken externally. In principle you can use any combination of day of week, year, minute, hour, etc..., all of which are common on all nodes.

5102

General Questions and Answers / Re: How to set up the parameters to "Geometric Constraints" in VNL?

« on: May 25, 2009, 16:09 »

The electrode atoms cannot be moved during a relaxation, so if all atoms in the central region will be allowed to move, you do not need to set up any constraints. However, it is often a very good idea to keep at least the first layer or two of the surface region (the part of the central region which is like the electrodes) fixed to have a better match to the electrodes, and to reduce the computation time.

5103

General Questions and Answers / Re: Speed issue in Two probe configuration

« on: May 25, 2009, 15:26 »

It's kind of hard to troubleshoot such things remotely... Is the difference reproducible each time you run? No other jobs running that might load the nodes with other jobs?

You write threads, but I assume you mean MPI nodes, i.e. on the 8-node cluster you run mpiexec with "-n 8", and "-n 16" on the 16-node cluster?

5104

Scripts, Tutorials and Applications / Re: Script for calculate the bandgap.

« on: May 25, 2009, 09:13 »

The way the picture looks, this system is metallic and hence there is no band gap. The script assumes a band gap exists; if it doesn't you may indeed get weird results. This is a good example of the importance of not just applying a script like this and trust the results blindly, but one must always investigate the results in detail.

5105

General Questions and Answers / Re: question about calculate the local_density_of_states

« on: May 25, 2009, 09:05 »

Sure, it just requires a bit of manual scripting. Use the toArray() function to extract the data as a Numpy array, then you can manipulate the data as you want.

For a similar example, see http://quantumwise.com/forum/index.php?topic=21.0; the script voltagedrop.py in first post there shows how to extract the data and do some simple manipulations. Also see http://quantumwise.com/forum/index.php?topic=39.0.

5106

General Questions and Answers / Re: Choice of basis

« on: May 24, 2009, 22:34 »

A concrete example should make it clear.

Hydrogen
Valence: 1s
SZ: s
DZ: s s'
SZP: s p
DZP: s s' p
DZDP: s s' p p'

Gold
Valence: 6s 5d^10
SZ: s d
DZ: s s' d d'
SZP: s d p
DZP: s s' d d' p
DZP: s s' d d' p p'

By s' we here mean a second s orbital (same major quantum number) to add more radial degrees of freedom; the polarization functions add angular degrees of freedom.

The major quantum number are excluded since the basis orbitals are based on the pseudocore atom. Thus, the 6s valence electrons in the basis set for gold really look like 1s wavefunctions.

For a list of the valence as used by ATK, see http://quantumwise.com/documents/manuals/ATK-2008.10/ref.atomdata.html.

5107

General Questions and Answers / Re: question about calculate the local_density_of_states

« on: May 24, 2009, 19:25 »

My guess is that you used VNL to generate the script, and ticked the box to restore self-consistent calculation from checkpoint. In that case, to proceed with analysis and calculating the LDOS etc, you must also remember to untick the box "Only use initial density" (which is used when you wish to restart a calculation).

5108

General Questions and Answers / Re: two questions on "method" setup in VNL

« on: May 24, 2009, 19:21 »

Confusing as it may be, the "charge" in the basis set parameters has nothing to do with the charge of the atom. It's just a parameter used to generate the basis set. All atoms are assumed to be neutral in ATK, but charge transfer can (and will) occur in a two-probe system.

5109

General Questions and Answers / Re: question about calculate the local_density_of_states

« on: May 24, 2009, 11:10 »

This error does not arise from the lines of code you have quoted, but somewhere earlier (probably) in the script. Most likely you are trying to initialize a new calculation based on an existing checkpoint file (nc file). This only works if they systems are identical in terms of atoms, their order, and the basis set.

The notation for the quantum numbers can be a bit tricky. For the example you show it's a single k-point, at the Fermi energy (given separately) for spin down.

5110

General Questions and Answers / Re: How to set the Quantum numbers when calculating the MPSH?

« on: May 22, 2009, 22:40 »

The VNL checkboxes are really just a quick way to select certain atoms (which, anyway, doesn't work unless the structure conforms to the surface/molecule/surface model). The wavefunctions are always indexed to match the eigenvalues (the spectrum), starting at 0.

5111

General Questions and Answers / Re: Speed issue in Two probe configuration

« on: May 22, 2009, 18:29 »

Do you use the exact same parallelization strategy on the two cluster? I'm particularly referring to the loading of MPI nodes vs. CPUs w.r.t. the number of cores/sockets. You should avoid using more MPI nodes than physical nodes (i.e., if you have for instance 2 dual cores, you may not get very good performance using 4 MPI nodes on this system).

Also, sometimes on Linux the OS will not show the correct number of cores to thread on to the application. You can ensure ATK knows how many threads it can start on each node by setting MKL_NUM_THREADS by hand (see the manual!).

5112

General Questions and Answers / Re: Choice of basis

« on: May 22, 2009, 17:44 »

There is a fundamental difference between TB and the orbitals that are used in e.g. DFT. In TB you focus on the bonds between neighboring atoms, while the orbitals used in ATK are used to expand the electron density in the vicinity of each atom. In this case, the "bonding" comes about from overlap of such orbitals between atoms within a certain range. In ATK we use orbitals with a finite range (hard cutoff), and atoms within this range have a chance to "bond" in this way. The range is independent of the basis set type, but can be controlled by the "energy shift", as described in the manual.

For more information, see also http://quantumwise.com/documents/manuals/ATK-2008.10/ref.atomicorbitals.html.

5113

Scripts, Tutorials and Applications / Re: Script for calculate the bandgap.

« on: May 15, 2009, 13:16 »

The value for (7,0) seems fine, while the negative value for (6,0) probably is due to the fact that it's semi-metallic.

I suggest you look at a plot of band structure and try to relate it to the numbers you get. If you attach those plots (properly zoomed in around the Fermi energy), we can understand the situation better.

5114

General Questions and Answers / Re: question about Li-H2-Li two-probe system

« on: May 14, 2009, 08:38 »

Most likely the calculation has converged to a state with zero charge in the entire system. Au is trickier than Li and requires not only a bit higher temperature but most likely also a longer electrode.

5115

General Questions and Answers / Re: question on two-probe model setting up

« on: May 13, 2009, 13:37 »

It seems your pictures did not get attached to the post.

When it comes to selecting the Li atom, you will not see the actual atom get selected, that's true (and a bit confusing). You just need to have the mouse over it when you right-click to choose translate. Hope this helps.