1 GROMACS 2016.5 Release Notes
2 ----------------------------
4 This version was released on February 16, 2018. These release notes
5 document the changes that have taken place in GROMACS since version
6 2016.4 to fix known issues. It also incorporates all fixes made in
7 version 5.1.5 (which was the last planned release in that series).
9 Fixes where mdrun could behave incorrectly
10 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
12 Fixed triclinic domain decomposition bug
13 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
14 With triclinic unit-cells with vectors a,b,c, the domain decomposition
15 would communicate an incorrect halo along dimension x when b[x]!=0
16 and vector c not parallel to the z-axis. The halo cut-off bound plane
17 was tilted incorrect along x/z with an error approximately
18 proportional to b[x]*(c[x] - b[x]*c[y]/b[y]).
19 When c[x] > b[x]*c[y]/b[y], the communicated halo was too small, which
20 could cause instabilities or silent errors.
21 When c[x] < b[x]*c[y]/b[y], the communicated halo was too large, which
22 could cause some communication overhead.
26 Required -ntmpi with setting -ntomp with GPUs
27 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
28 With GPUs and thread-MPI, setting only ``-ntomp`` could lead to
29 oversubscription of the hardware threads.
30 Now with GPUs and thread-MPI the user is required to set ``-ntmpi`` when
31 using ``-ntomp``. Here we chose that to also require ``-ntmpi`` when the user
32 specified both ``-nt`` and ``-ntomp``; here we could infer the number of
33 ranks, but it's safer to ask the user to explicity set ``-ntmpi``.
34 Note that specifying both ``-ntmpi`` and ``-nt`` has always worked correctly.
38 Prevented dynamic load balancing activating immediately after exchange
39 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
40 Turning on DLB right after exchanging replicas caused an assertion
41 failure and is also useless.
45 Avoided confusing message at end of non-dynamical runs
46 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
47 EM, TPI, NM, etc. are not targets for performance optimization
48 so we will not write performance reports. This commit fixes
49 and oversight whereby we would warn a user when the lack of
50 performance report is normal and expected.
54 Changed to issue fewer messages when ``-cpi`` checkpoint file is not present
55 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
56 Removed duplicated message.
60 Disallowed combination of PME-user and Verlet cutoff
61 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
64 Added missing Ewald correction for pme-user
65 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
66 With ``coulomb-type = pme-user``, the Ewald mesh energy was not subtracted
67 leading to (very) incorrect Coulomb energies and forces.
71 Fixed thread-MPI rank choice for orientation restraints
72 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
73 Only a single rank is supported, so that must be what the thread-MPI
74 code will choose. There's another check later on that catches the
77 Fixed nstlist increase warning print
78 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
79 The log file warning message had a buggy conditional which this commit
82 Removed incorrect comment for CHARMM tips3p
83 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
85 Fixed incorrect dV/dlambda for walls
86 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
87 The free-energy derivative dV/dlambda for walls, which can
88 be perturbed by changing atom types of non-wall atoms, only
89 contained the B-state contribution.
93 Fixed warning for confout with periodic molecules
94 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
95 With periodic molecules, ``gmx mdrun`` would incorrectly attempt to make
96 molecules whole for writing the final state to confout.
100 Fixed wrong megaflop accounting
101 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
102 Some nrnb index entries were missing in the interaction_function
103 array, leading to that wrong megaflops accounting printed.
107 Fixes for ``gmx`` tools
108 ^^^^^^^^^^^^^^^^^^^^^^^
110 Fixed ``gmx grompp`` net charge check
111 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
112 The grompp check for the net charge would ignore molecule blocks
113 at the end when molecule types are used in multiple, non consecutive
118 Extended ``gmx grompp`` missing energy term message
119 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
122 Fixed ``gmx genion`` charge summation accuracy
123 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
124 ``gmx genion`` accumulated the charge is a float, which could cause
125 underestimation of the net charge for highly charged systems.
129 Fixed ``gmx check`` for tprs with different #atoms
130 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
133 Fixed ``gmx grompp`` with Andersen massive and no COM removal
134 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
135 Fixed a floating point exception leading to a crash.
136 Also fixed possible different rounding for the interval for
137 Andersen massive in ``gmx grompp`` from ``gmx mdrun`` for
138 the common case where ``tau_t`` is a multiple of ``delta_t``.
142 Improved documentation
143 ^^^^^^^^^^^^^^^^^^^^^^
145 Updated documention of Nose-Hoover output
146 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
147 The documentation of Nose-Hoover chain variable printing was
152 Clarified docs for Fmax in EM
153 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""