Bugs fixed#
Fixed exported libgromacs CMake target#
Update the exported libgromacs CMake target to not depend on non- existing include paths and add GMX_DOUBLE define to interface definitions. The target now gets exported into the Gromacs namespace.
Fixed unsolicited changing of atom names in pdb file#
Remove functions to change atoms names when reading and writing pdb files. This affected naming of H atoms in particular.
pdb2gmx handles ASPH and GLUH inputs better#
The default is to treat all such residues as the unprotonated form, and not to try to infer the protonation from the residue name in the input. Protonated forms are only available via the interactive selection options. Now pdb2gmx reports when it is converting such input residues automatically. It also ensures that the output configuration and topology are naming such residues correctly in both the default and interactive-selection cases.
Correct excluded perturbed interactions beyond the non-bonded cut-off distance#
With free-energy calculations without coupling of intermolecular interactions, non-bonded pair interactions at distance longer than the cut-off distance can be excluded. These interactions would still have PME long-range contributions. The contributions are now removed. In addition, mdrun will stop with a fatal error when interactions beyond the pair-list cut-off are present.
Corrected AWH initial histogram size#
The initial histogram size for AWH biases depended (weakly) on the force constant. This dependence has been removed, which increases the histogram size by a about a factor of 3. In practice this has only a minor effect on the time to solution. For multiple dimensions, the histogram size was underestimated, in particular with a combination of slower and faster dimensions. The, now simplified, formula for the initial histogram size is given in the reference manual.
Fixed default for tick-mark spacing in gmx xpm2ps#
This was inadvertently changed many years ago, leading to the intended default of automatic tick-mark spacing being replaced with an unsuitable fixed value.
Fixed LJ Ewald exclusions when used with cut-off electrostatics#
The exclusion forces in CUDA and OpenCL kernels were computed incorrectly if LJ Ewald was used together with cut-off electrostatics.