GROMACS 2018.5 release notes ---------------------------- This version was released on January 22, 2019. These release notes document the changes that have taken place in GROMACS since version 2018.4, to fix known issues. A complete list of fixes and their descriptions can be found in the :ref:`release-notes`. Fixes where mdrun could behave incorrectly ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Fixed numerical derivative for normal-mode analysis with shell code """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" Due to higher precision needs when using polarizable shell particles the normal mode code did not work reproducibly with shells. In order to fix this the step-size used for numerically computing the Hessian was reduced to near machine precision. The change does not affect the results for non-polarizable systems, such as proteins or small molecules. Make large PME grids work on GPU """"""""""""""""""""""""""""""""""""""""""" PME grids with size along Z larger than 511 would make mdrun exit with a cryptic CUDA error. :issue:`2779` Fix LINCS accuracy with OpenMP when constraint triangles are present """""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" Constraint triangles, which usually only occur when replacing hydrogens by virtual interaction sites in CH3 and NH3 groups, need double the number of iterations as normal constraints. With OpenMP this would only happen when the last OpenMP thread has at least one such triangle. This would cause a slight loss of accuracy in inhomogeneous systems. :issue:`2808` Fix acceleration with ``cos-acceleration`` """""""""""""""""""""""""""""""""""""""""" A factor of 2 was missing from the acceleration value, leading to incorrect results when e.g. calculating viscosities. :issue:`2572` Fix checkpoint restart of tpr with infinite step count """""""""""""""""""""""""""""""""""""""""""""""""""""" An issue was introduced that caused :ref:`mdrun <gmx mdrun>` to refuse to start when using infinite step counts. Now :ref:`mdrun <gmx mdrun>` properly accepts those files again. :issue:`2757` Fix energy history file reading """"""""""""""""""""""""""""""" A check was missing when reading energy files from a simulation that was stopped before the first value had been written to the file. :issue:`2781` Fixes for ``gmx`` tools ^^^^^^^^^^^^^^^^^^^^^^^ Fix trjconv -ndec """"""""""""""""""""""""""""""""""""""""""""""""""""""""" This only works for writing .xtc files. The code and documentation now works correctly with .gro files, which was changed in 2016 release series so that it would only write fixed-width columns. :issue:`2813` :issue:`2037` Fixes to improve portability ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Miscellaneous ^^^^^^^^^^^^^ Warn for problematic coupling times with Nose-Hoover and Parrinello-Rahman """""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" When combining the Nose-Hoover and Parrinello-Rahman coupling algorithms, resonances in the kinetic energy and pressure/volume can appear when the two coupling times involved are similar. Now grompp warns when ``tau-p`` is less than two times ``tau-t``. :issue:`2749` Fixed efficiency issue with shell code minimization """""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" Code cleanup touching unnecessarily complex code created an efficiency issue. Both the issue and some of the complexity are now fixed. :issue:`2705` Added code generation support for NVIDIA Turing GPUs """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" With CUDA 10.0 NVIDIA Turing GPUs can be directly targeted by the nvcc compiler. We now generate the appropriate set of flags for the Turing architecture by default when using CUDA 10 (or later).