DFTB+ 1.0.1

This is a maintenance release for version 1.0. No new features had been added.

Compilation

DFTB+ 1.0.1 had been successfully compiled and tested (using the tests in the autotest system) on the following architectures:

Machine	System	Compiler	BLAS/LAPACK	Notes
i686	Linux	Intel Fortran (32-bit) 9.1	MKL 8.0
i686	Linux	Intel Fortran (32-bit) 10.0	MKL 9.1
i686	Linux	g95 (build 20060727)	ATLAS
x86_64	Linux	Intel Fortran(em64t) 9.1	MKL 8.0
x86_64	Linux	Intel Fortran (em64t) 10.0	MKL 9.1

 Changes

Since version 1.0 the following changes had been made:

Make process does not use Perl any more.

Due to some issues with broken Perl versions which caused the DFTB+ compilation process to fail, we omitted the usage of Perl during compilation.

Velocities in the xyz output corrected.

Due to an error in the conversion factors, the velocities written to the xyz-files (in units AA/ps) were not correct. Please note, that this only affected the output but not the internal behaviour of the code. Internally the code uses atomic units for all quantities without the need for any conversion.

Program reported the system to be neutral even if Charges had been set to nonzero.

The program reported the initial charges to be set to neutral, even if the number of charges had been set to a non-zero value by using the Charges keyword. The bug affected only the output, internally the correct number of charges was used (as correctly reported in the file detailed.out).

Electron counting for special cases fixed.

The number of electrons in the system was calculated incorrectly if Slater-Koster files with occupied d-orbitals were used, but only a maximal angular momentum of "p" was specified for the corresponding element in the input. (Apart from testing, the highest angular momentum specified in the input should be the same, as the one used when creating the corresponding SK-files.)

Weight for the k-points in supercell folding fixed

The weight for the Gamma-point was calculated incorrectly by the code, if following criteria were met:

• The K-points for a calculation were generated using the SupercellFolding method
• The K-point set (before the automatic reduction) was inversion symmetric
• the K-point set contains more than one point
• the K-point set contains the Gamma-point

In those cases the calculated weight of the Gamma point was double the correct value. This shouldn't change affect your results significantly, unless you used small odd Monkhorst-Pack K-point sets.

Gen format reader more robust

Program reports error, if the GenFormat tag contains inconsistent data (e.g. too many lines of data).

Changing a few typos in the manual