ATK 13.8.1 update
Feb 20 2014
An updated version of ATK 13.8 is available for download. A few somewhat serious bugs have been fixed, in addition to several minor ones, and there are also some new features!
This release requires license version 13.8 or higher; in short, if you could run 13.8.0 you can (and should!) update to this release.
For information on the new features in ATK 13.8 compared to older versions, see the ATK 13.8.0 release letter!
New in ATK 13.8.1
Serious bugs fixed (which could give incorrect results in calculations!)
- The wrong set of k-points for hexagonal (and other) lattices were generated by the symmetry recognition routines in ATK in some cases, which could result in incorrect results for the density of states. In 13.8.1 this is solved by not using symmetries at all - this makes the calculations take a bit longer of course. In 14.2 the correct symmetry points will be used.
- The Cleaver was unable to cleave for instance base-centered orthorhombic crystals correctly. Also, there were issues when you manually rotated the cell - even if you chose an "electrode" cell, C was not perpendicular to the AB plane always.
- The dynamical matrix calculation now employs the acoustic sum rule and symmetries to avoid negative frequencies.
- A degeneracy factor was missing for the entropy of phonon DOS.
- The function sortCoordinates (used e.g. in the "Coordinate list" plugin) would fail for systems with many atoms, leaving the atoms unsorted. This was especially a problem for large devices which rely on this sorting for the central region to reduce memory usage (and indeed construct the device using "Device from Bulk").
- A Bohr/Angstrom mixup in QuantumEspresso import caused incorrect structures if the input file was specified in Bohr.
- Errors with Tersoff potentials for III-V alloys (Tersoff_GaAs_2002, Tersoff_GaAs_2011 and Tersoff_InGaAs_2000) have been corrected (the errors are actually in the original references themselves).
Additions & improvements
- LDOS for ATK-SE has been implemented.
- Thermal transport coefficients can now be plotted as a function of the Fermi level (image to the right)
- ElectrodeValidator function - new way to find proper electrode sizes (will be presented in the new ATK device tutorial)
- Adding more information to the About box, to help diagnose OpenGL issues. Also possible to email the information directly to QuantumWise.
- ElectronDifferenceDensity now calculates all spin components.
- All files are unchecked by default in new projects, to avoid a large new project taking very long time to load.
- The dynamical matrix is no longer recalculated for devices each time - now it will be reused if it has already been calculated.
- The colorbar is now included in exported images.
- Export Abinit scripts to a directory without actually running them (this is not supported from the Script Generator though, you have to add a line in the script).
- Band structure plots etc now have window titles so you know which file they come from.
- The object ID is shown in the tooltips for items imported to the LabFloor from NC files.
- You can now delete files using the keyboard (Del button) in the file panel of the main VNL window.
- The Installation guide has been added to the Start Menu.
- Sorting of the projects in the "Open project" dialog has been disabled because sorting caused the wrong project to be opened when you selected it.
- LabFloor importers now give item titles to FHI and QuantumEspresso files.
- The scrollbars on the LabFloor work better now.
Noticeable but not severe bugs
- It was not possible to select multiple images in a NEB path in the Builder - and thus not possible to apply operations like Translate etc to many images at once.
- The Script Generator made incorrect scripts for GGA and MGGA with ABINIT.
- Spin-polarized device calculations using DFTB now works - the equivalent bulk was not polarized which caused a segfault when going into the device part.
- When exporting matplotlib data, the global normalization factor was not removed, so the data was not scaled properly.
- Running MemoryUsage for device using a DFTB calculator now works.
- Cut Planes in the Viewer would not display negative values.
- Running a quick optimization with the Brenner potential made it impossible to insert a spatial region.
- LDOS was not supported in the Projector1D plugin
- Crystal Builder: if the first inserted point is (x,y,z), editing coordinates didn't work.
- The NEB builder progress bar appeared behind the Builder and the "Create" button could easily be clicked multiple times inf you didn't notice it was already running.
- Cut Planes in the Viewer - the preset planes AC and BC were reversed.
- A smaller Bohr/Angstrom mixup for unit cells has been fixed.
- "Analysis from File" (and a few other file dialogs) did not always open up in the project directory.
- Crystal Builder: the default unique axis for Monoclinic should be B
- The discs at the end of bonds were not transparent.
- filename.nc was written twice when LabFloor items were grouped by calculator ID
- Classic mode started in the wrong directory sometimes.
- Copy atoms is no longer possible for NEB configurations in plugins like Translate etc (it doesn't make sense).
Unfixed known issues
These will not be fixed until the next major release of ATK, but we are aware of them.
- Running Abinit in parallel requires special setup of the cluster, and even with that it sometimes doesn't work properly. We are investigating the issue but don't expect to have a fix until ATK 14.2.
- Sometimes you click the "Create" button in the I-V Curve generator but nothing happens. Solution: try again! It's because of a conflict with the file being locked for reading by the LabFloor, and the file is unavailable for writing a short period after that.
- Coloring of isosurfaces for Bloch states is wrong.
- Reloading plugins raises errors if any atoms are selected when the reload is requested. Workaround: unselect all atoms first.
- Please report any additional isses at .