1GROMACS 2019.4 release notes
2----------------------------
3
4This version was released on October 2nd, 2019. These release notes
5document the changes that have taken place in GROMACS since the
6previous 2019.3 version, to fix known issues. It also incorporates all
7fixes made in version 2018.7 and earlier, which you can find described
8in the :ref:`release-notes`.
9
10.. Note to developers!
11   Please use """"""" to underline the individual entries for fixed issues in the subfolders,
12   otherwise the formatting on the webpage is messed up.
13   Also, please use the syntax :issue:`number` to reference issues on GitLab, without the
14   a space between the colon and number!
15
16Fixes where mdrun could behave incorrectly
17^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
18
19Fix incorrect pressure when atoms in CMAP cross a box boundary
20""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
21
22The virial calculation and thus the pressure would be incorrect
23when the second and third atom involved in a CHARMM CMAP correction
24term would reside in different periodic images. This can happen when
25a protein is positioned over a box boundary. Note that the energy
26and forces were correct, but sampling was affected when pressure
27coupling was applied when a protein crossed a box boundary.
28
29:issue:`2845`
30:issue:`2867`
31
32Fix incorrect LJ cut-off on GPU when rvdw < rcoulomb
33""""""""""""""""""""""""""""""""""""""""""""""""""""
34
35When rvdw was chosen by the user to be smaller than rcoulomb in the mdp file,
36the LJ cut-off would initially be set to the Coulomb cut-off for computing
37non-bonded interactions on the GPU. This only affected energy minimization,
38mdrun -rerun and the first 2*nstlist steps of a normal MD run, since the correct
39LJ cut-off is set when PME tuning (on by default) starts after 2*nstlist steps
40(unless PME tuning was disabled with -notunepme).
41
42:issue:`3056`
43
44
45Fix (unlikely) missing bonded forces with CUDA GPUs and domain decomposition
46""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
47
48Forces could be missing for bonded interactions computed on CUDA GPUs with
49domain decomposition when there are non-local bonded interactions, but no
50non-local non-bonded interactions between two domains. Note that this is
51extremely unlikely to happen, since the distance between the bonded atoms
52needs to be larger than the pair-list cut-off distance and there should be no
53other non-local atoms within the pair-list cut-off distance.
54
55:issue:`3063`
56
57Fix incorrect reporting of final kinetic energy and temperature
58"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
59
60With the leap-frog integrator, the kinetic energy and temperature reported
61for the last step were incorrect when the last step was not divisible by
62nstcalcenergy, nsttcouple or nstpcouple.
63
64:issue:`2950`
65
66Fix segmentation fault in grompp and mdrun with cosine COM pulling
67""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
68
69:issue:`3023`
70
71
72Fixes for ``gmx`` tools
73^^^^^^^^^^^^^^^^^^^^^^^
74
75Fix grompp not adding angle constraints between constraints
76"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
77
78When using the mdp option constraints=all-angles, angles involving
79bonds supplied as constraints in the topology would be removed,
80but not replaced by angle constraints.
81
82:issue:`3067`
83
84Fix gmx wham with angle and dihedral geometries
85"""""""""""""""""""""""""""""""""""""""""""""""
86
87gmx wham would apply an incorrect radian to degree unit conversion,
88leading to no overlap or not-a-number output.
89
90:issue:`2609`
91:issue:`3094`
92
93Fix bug in gmx xpm2ps
94"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
95
96The tool would fail when not being provided with a library file to read in.
97
98:issue:`3012`
99
100Fix bug in gmx anaeig
101"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
102
103An issue was noted when reading a second set
104set of eigenvectors that could lead to problems when the number
105of eigenvectors was less than the three times the number of atoms.
106
107:issue:`2972`
108
109Fix issue with demux.pl script
110"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
111
112The trajectories could become discontinuous with simulations longer than 100ns
113and exchange strides that are not a multiple of 1 ps. This only affected the
114post-processing of trajectories generated from replica exchange simulations.
115
116Made gmx disre work with non-consecutively labeled restraints
117"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
118
119:issue:`2953`
120
121Fixed writing of gro files with index groups
122"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
123
124An output ``.gro`` file from from e.g. ``gmx editconf -f -n`` would
125take the atom names for the output file in order from the atoms in the
126input file, rather than in order from the atoms indicated by the
127indices in the index file.
128
129:issue:`3107`
130
131Made gmx make_ndx keep chain IDs
132"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
133
134Old style structure file reading caused the chain IDs to be overwritten with
135default values.
136
137:issue:`3070`
138
139Fixes that affect portability
140^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
141
142Disable PME OpenCL on Apple
143"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
144
145The Apple OpenCL compilers fail to produce a functional clFFT build.
146The OpenCL PME support is therefore disabled on Apple platforms.
147
148:issue:`2941`
149
150Miscellaneous
151^^^^^^^^^^^^^
152
153Added AMD Zen 2 detection
154^^^^^^^^^^^^^^^^^^^^^^^^^
155
156The AMD Zen 2 architecture is now detected as different from Zen 1
157and uses 256-bit wide AVX2 SIMD instructions (GMX_SIMD=AVX2_256) by default.
158Also the non-bonded kernel parameters have been tuned for Zen 2.
159This has a significant impact on performance.
160