As of ATK 11.8, we have started working on a rather general way to set up calculations in other codes than ATK from our graphical user interface Virtual NanoLab (VNL). We already have prototypes for VASP and a few other codes.
For users who until now have used VNL to set up GPAW
calculations, there are few changes in version 11.8. Since we unfortunately have not had time to update the tutorial to reflect these changes, let me outline the changes below.
The most immediate change is that we have included a brand new GPAW Scripter in 11.8. However, it is no longer located in the Script Generator, as it was before. To find the new GPAW interface, go to the Tools
menu in VNL (the main window) and choose Custom Scripter
. In there, go to the menu Scripters
and choose GPAW Scripter
. Hopefully it's easy to work with; to get an atomic configuration into it, just drop the configuration from the Builder or Database etc onto the Scripter window (you can drop it anywhere).
This is still an early version of the GPAW scripter, thus there are some limitations. But at least you can build structures, using the other tools in VNL (and, most notably, in 12.2 coming out soon, there will be a completely new and radically improved Builder!), and get a basic GPAW input script with very little effort. Then you can edit the script manually if there are some calculations you want to do which our Scripter does not support yet.
The second part of the change will be that we will no longer ship a precompiled GPAW with the ATK package; this will take effect from version 12.2. This means a somewhat more involved procedure for the user, who needs to download and compile GPAW (unless you are able to use one of their precompiled versions), but one should note that the GPAW shipped with ATK 11.8 and earlier is
a) an older version (0.6.5147, while the latest official GPAW is already 0.9), and
b) only capable of running in serial - if you know a bit about GPAW, you will realize that the whole point of the code is to run in parallel (or else performance suffers a lot).
For now, you can still in principle use our 0.6 version, but I wouldn't recommend it, esp. since I know there were a few bugs back in 2009, which should have been fixed by now in 0.8.
Our plan is to extend the support for GPAW quite a lot in the near future. Thus we value any feedback users can provide on the current Scripter, so we know how to improve it - and VNL in general.
For more information about GPAW, and to download the latest version, please go to https://wiki.fysik.dtu.dk/gpaw/