start mos without define

[tmartin]

I have a bunch of jobs, where the DFT single point calculations don't converge. These jobs are part of a PES and the starting mos originate from the global minimum, but seem to be a poor guess for geometries too different. After redoing all the define input, the calculations converge. Now I search for a method to batch all this.

Since piping input into define is a pain in the ass (sometimes it produces segmentation faults without throwing an error), is it possible to generate start mos without using define?

[Arnim]

Hi,

you can run 'hcore' to start from a Hamilton core guess. It might take longer to converge, though.

Cheers,

Arnim