2 See the "run control" section for a working example of the
3 syntax to use when making .mdp entries, with and without detailed
4 documentation for values those entries might take. Everything can
5 be cross-referenced, see the examples there. TODO Make more
8 Molecular dynamics parameters (.mdp options)
9 ============================================
14 Default values are given in parentheses, or listed first among
15 choices. The first option in the list is always the default
16 option. Units are given in square brackets The difference between a
17 dash and an underscore is ignored.
19 A :ref:`sample mdp file <mdp>` is available. This should be
20 appropriate to start a normal simulation. Edit it to suit your
21 specific needs and desires.
29 directories to include in your topology. Format:
30 ``-I/home/john/mylib -I../otherlib``
34 defines to pass to the preprocessor, default is no defines. You can
35 use any defines to control options in your customized topology
36 files. Options that act on existing :ref:`top` file mechanisms
39 ``-DFLEXIBLE`` will use flexible water instead of rigid water
40 into your topology, this can be useful for normal mode analysis.
42 ``-DPOSRES`` will trigger the inclusion of ``posre.itp`` into
43 your topology, used for implementing position restraints.
51 (Despite the name, this list includes algorithms that are not
52 actually integrators over time. :mdp-value:`integrator=steep` and
53 all entries following it are in this category)
57 A leap-frog algorithm for integrating Newton's equations of motion.
61 A velocity Verlet algorithm for integrating Newton's equations
62 of motion. For constant NVE simulations started from
63 corresponding points in the same trajectory, the trajectories
64 are analytically, but not binary, identical to the
65 :mdp-value:`integrator=md` leap-frog integrator. The the kinetic
66 energy, which is determined from the whole step velocities and
67 is therefore slightly too high. The advantage of this integrator
68 is more accurate, reversible Nose-Hoover and Parrinello-Rahman
69 coupling integration based on Trotter expansion, as well as
70 (slightly too small) full step velocity output. This all comes
71 at the cost off extra computation, especially with constraints
72 and extra communication in parallel. Note that for nearly all
73 production simulations the :mdp-value:`integrator=md` integrator
76 .. mdp-value:: md-vv-avek
78 A velocity Verlet algorithm identical to
79 :mdp-value:`integrator=md-vv`, except that the kinetic energy is
80 determined as the average of the two half step kinetic energies
81 as in the :mdp-value:`integrator=md` integrator, and this thus
82 more accurate. With Nose-Hoover and/or Parrinello-Rahman
83 coupling this comes with a slight increase in computational
88 An accurate and efficient leap-frog stochastic dynamics
89 integrator. With constraints, coordinates needs to be
90 constrained twice per integration step. Depending on the
91 computational cost of the force calculation, this can take a
92 significant part of the simulation time. The temperature for one
93 or more groups of atoms (:mdp:`tc-grps`) is set with
94 :mdp:`ref-t`, the inverse friction constant for each group is
95 set with :mdp:`tau-t`. The parameter :mdp:`tcoupl` is
96 ignored. The random generator is initialized with
97 :mdp:`ld-seed`. When used as a thermostat, an appropriate value
98 for :mdp:`tau-t` is 2 ps, since this results in a friction that
99 is lower than the internal friction of water, while it is high
100 enough to remove excess heat NOTE: temperature deviations decay
101 twice as fast as with a Berendsen thermostat with the same
106 This used to be the default sd integrator, but is now
107 deprecated. Four Gaussian random numbers are required per
108 coordinate per step. With constraints, the temperature will be
113 An Euler integrator for Brownian or position Langevin dynamics,
114 the velocity is the force divided by a friction coefficient
115 (:mdp:`bd-fric`) plus random thermal noise (:mdp:`ref-t`). When
116 :mdp:`bd-fric` is 0, the friction coefficient for each particle
117 is calculated as mass/ :mdp:`tau-t`, as for the integrator
118 :mdp-value:`integrator=sd`. The random generator is initialized
123 A steepest descent algorithm for energy minimization. The
124 maximum step size is :mdp:`emstep`, the tolerance is
129 A conjugate gradient algorithm for energy minimization, the
130 tolerance is :mdp:`emtol`. CG is more efficient when a steepest
131 descent step is done every once in a while, this is determined
132 by :mdp:`nstcgsteep`. For a minimization prior to a normal mode
133 analysis, which requires a very high accuracy, |Gromacs| should be
134 compiled in double precision.
136 .. mdp-value:: l-bfgs
138 A quasi-Newtonian algorithm for energy minimization according to
139 the low-memory Broyden-Fletcher-Goldfarb-Shanno approach. In
140 practice this seems to converge faster than Conjugate Gradients,
141 but due to the correction steps necessary it is not (yet)
146 Normal mode analysis is performed on the structure in the :ref:`tpr`
147 file. |Gromacs| should be compiled in double precision.
151 Test particle insertion. The last molecule in the topology is
152 the test particle. A trajectory must be provided to ``mdrun
153 -rerun``. This trajectory should not contain the molecule to be
154 inserted. Insertions are performed :mdp:`nsteps` times in each
155 frame at random locations and with random orientiations of the
156 molecule. When :mdp:`nstlist` is larger than one,
157 :mdp:`nstlist` insertions are performed in a sphere with radius
158 :mdp:`rtpi` around a the same random location using the same
159 neighborlist (and the same long-range energy when :mdp:`rvdw`
160 or :mdp:`rcoulomb` > :mdp:`rlist`, which is only allowed for
161 single-atom molecules). Since neighborlist construction is
162 expensive, one can perform several extra insertions with the
163 same list almost for free. The random seed is set with
164 :mdp:`ld-seed`. The temperature for the Boltzmann weighting is
165 set with :mdp:`ref-t`, this should match the temperature of the
166 simulation of the original trajectory. Dispersion correction is
167 implemented correctly for TPI. All relevant quantities are
168 written to the file specified with ``mdrun -tpi``. The
169 distribution of insertion energies is written to the file
170 specified with ``mdrun -tpid``. No trajectory or energy file is
171 written. Parallel TPI gives identical results to single-node
172 TPI. For charged molecules, using PME with a fine grid is most
173 accurate and also efficient, since the potential in the system
174 only needs to be calculated once per frame.
178 Test particle insertion into a predefined cavity location. The
179 procedure is the same as for :mdp-value:`integrator=tpi`, except
180 that one coordinate extra is read from the trajectory, which is
181 used as the insertion location. The molecule to be inserted
182 should be centered at 0,0,0. |Gromacs| does not do this for you,
183 since for different situations a different way of centering
184 might be optimal. Also :mdp:`rtpi` sets the radius for the
185 sphere around this location. Neighbor searching is done only
186 once per frame, :mdp:`nstlist` is not used. Parallel
187 :mdp-value:`integrator=tpic` gives identical results to
188 single-rank :mdp-value:`integrator=tpic`.
193 starting time for your run (only makes sense for time-based
199 time step for integration (only makes sense for time-based
205 maximum number of steps to integrate or minimize, -1 is no
211 The starting step. The time at an step i in a run is
212 calculated as: t = :mdp:`tinit` + :mdp:`dt` *
213 (:mdp:`init-step` + i). The free-energy lambda is calculated
214 as: lambda = :mdp:`init-lambda` + :mdp:`delta-lambda` *
215 (:mdp:`init-step` + i). Also non-equilibrium MD parameters can
216 depend on the step number. Thus for exact restarts or redoing
217 part of a run it might be necessary to set :mdp:`init-step` to
218 the step number of the restart frame. :ref:`gmx convert-tpr`
219 does this automatically.
223 .. mdp-value:: Linear
225 Remove center of mass translation
227 .. mdp-value:: Angular
229 Remove center of mass translation and rotation around the center of mass
233 No restriction on the center of mass motion
238 frequency for center of mass motion removal
242 group(s) for center of mass motion removal, default is the whole
252 Brownian dynamics friction coefficient. When :mdp:`bd-fric` is 0,
253 the friction coefficient for each particle is calculated as mass/
259 used to initialize random generator for thermal noise for
260 stochastic and Brownian dynamics. When :mdp:`ld-seed` is set to -1,
261 a pseudo random seed is used. When running BD or SD on multiple
262 processors, each processor uses a seed equal to :mdp:`ld-seed` plus
263 the processor number.
271 (10.0) \[kJ mol-1 nm-1\]
272 the minimization is converged when the maximum force is smaller
283 frequency of performing 1 steepest descent step while doing
284 conjugate gradient energy minimization.
289 Number of correction steps to use for L-BFGS minimization. A higher
290 number is (at least theoretically) more accurate, but slower.
293 Shell Molecular Dynamics
294 ^^^^^^^^^^^^^^^^^^^^^^^^
296 When shells or flexible constraints are present in the system the
297 positions of the shells and the lengths of the flexible constraints
298 are optimized at every time step until either the RMS force on the
299 shells and constraints is less than :mdp:`emtol`, or a maximum number
300 of iterations :mdp:`niter` has been reached. Minimization is converged
301 when the maximum force is smaller than :mdp:`emtol`. For shell MD this
302 value should be 1.0 at most.
307 maximum number of iterations for optimizing the shell positions and
308 the flexible constraints.
313 the step size for optimizing the flexible constraints. Should be
314 chosen as mu/(d2V/dq2) where mu is the reduced mass of two
315 particles in a flexible constraint and d2V/dq2 is the second
316 derivative of the potential in the constraint direction. Hopefully
317 this number does not differ too much between the flexible
318 constraints, as the number of iterations and thus the runtime is
319 very sensitive to fcstep. Try several values!
322 Test particle insertion
323 ^^^^^^^^^^^^^^^^^^^^^^^
328 the test particle insertion radius, see integrators
329 :mdp-value:`integrator=tpi` and :mdp-value:`integrator=tpic`
338 number of steps that elapse between writing coordinates to output
339 trajectory file, the last coordinates are always written
344 number of steps that elapse between writing velocities to output
345 trajectory, the last velocities are always written
350 number of steps that elapse between writing forces to output
356 number of steps that elapse between writing energies to the log
357 file, the last energies are always written
359 .. mdp:: nstcalcenergy
362 number of steps that elapse between calculating the energies, 0 is
363 never. This option is only relevant with dynamics. With a
364 twin-range cut-off setup :mdp:`nstcalcenergy` should be equal to
365 or a multiple of :mdp:`nstlist`. This option affects the
366 performance in parallel simulations, because calculating energies
367 requires global communication between all processes which can
368 become a bottleneck at high parallelization.
373 number of steps that else between writing energies to energy file,
374 the last energies are always written, should be a multiple of
375 :mdp:`nstcalcenergy`. Note that the exact sums and fluctuations
376 over all MD steps modulo :mdp:`nstcalcenergy` are stored in the
377 energy file, so :ref:`gmx energy` can report exact energy averages
378 and fluctuations also when :mdp:`nstenergy` > 1
380 .. mdp:: nstxout-compressed
383 number of steps that elapse between writing position coordinates
384 using lossy compression
386 .. mdp:: compressed-x-precision
389 precision with which to write to the compressed trajectory file
391 .. mdp:: compressed-x-grps
393 group(s) to write to the compressed trajectory file, by default the
394 whole system is written (if :mdp:`nstxout-compressed` > 0)
398 group(s) to write to energy file
404 .. mdp:: cutoff-scheme
406 .. mdp-value:: Verlet
408 Generate a pair list with buffering. The buffer size is
409 automatically set based on :mdp:`verlet-buffer-tolerance`,
410 unless this is set to -1, in which case :mdp:`rlist` will be
411 used. This option has an explicit, exact cut-off at :mdp:`rvdw`
412 equal to :mdp:`rcoulomb`. Currently only cut-off,
413 reaction-field, PME electrostatics and plain LJ are
414 supported. Some :ref:`gmx mdrun` functionality is not yet
415 supported with the :mdp:`Verlet` scheme, but :ref:`gmx grompp`
416 checks for this. Native GPU acceleration is only supported with
417 :mdp:`Verlet`. With GPU-accelerated PME or with separate PME
418 ranks, :ref:`gmx mdrun` will automatically tune the CPU/GPU load
419 balance by scaling :mdp:`rcoulomb` and the grid spacing. This
420 can be turned off with ``mdrun -notunepme``. :mdp:`Verlet` is
421 faster than :mdp:`group` when there is no water, or if
422 :mdp:`group` would use a pair-list buffer to conserve energy.
426 Generate a pair list for groups of atoms. These groups
427 correspond to the charge groups in the topology. This was the
428 only cut-off treatment scheme before version 4.6, and is
429 **deprecated in |gmx-version|**. There is no explicit buffering of
430 the pair list. This enables efficient force calculations for
431 water, but energy is only conserved when a buffer is explicitly
440 Frequency to update the neighbor list (and the long-range
441 forces, when using twin-range cut-offs). When this is 0, the
442 neighbor list is made only once. With energy minimization the
443 neighborlist will be updated for every energy evaluation when
444 :mdp:`nstlist` is greater than 0. With :mdp:`Verlet` and
445 :mdp:`verlet-buffer-tolerance` set, :mdp:`nstlist` is actually
446 a minimum value and :ref:`gmx mdrun` might increase it, unless
447 it is set to 1. With parallel simulations and/or non-bonded
448 force calculation on the GPU, a value of 20 or 40 often gives
449 the best performance. With :mdp:`group` and non-exact
450 cut-off's, :mdp:`nstlist` will affect the accuracy of your
451 simulation and it can not be chosen freely.
455 The neighbor list is only constructed once and never
456 updated. This is mainly useful for vacuum simulations in which
457 all particles see each other.
466 Controls the period between calculations of long-range forces when
467 using the group cut-off scheme.
471 Calculate the long-range forces every single step. This is
472 useful to have separate neighbor lists with buffers for
473 electrostatics and Van der Waals interactions, and in particular
474 it makes it possible to have the Van der Waals cutoff longer
475 than electrostatics (useful *e.g.* with PME). However, there is
476 no point in having identical long-range cutoffs for both
477 interaction forms and update them every step - then it will be
478 slightly faster to put everything in the short-range list.
482 Calculate the long-range forces every :mdp:`nstcalclr` steps
483 and use a multiple-time-step integrator to combine forces. This
484 can now be done more frequently than :mdp:`nstlist` since the
485 lists are stored, and it might be a good idea *e.g.* for Van der
486 Waals interactions that vary slower than electrostatics.
490 Calculate long-range forces on steps where neighbor searching is
491 performed. While this is the default value, you might want to
492 consider updating the long-range forces more frequently.
494 Note that twin-range force evaluation might be enabled
495 automatically by PP-PME load balancing. This is done in order to
496 maintain the chosen Van der Waals interaction radius even if the
497 load balancing is changing the electrostatics cutoff. If the
498 :ref:`mdp` file already specifies twin-range interactions (*e.g.* to
499 evaluate Lennard-Jones interactions with a longer cutoff than
500 the PME electrostatics every 2-3 steps), the load balancing will
501 have also a small effect on Lennard-Jones, since the short-range
502 cutoff (inside which forces are evaluated every step) is
509 Make a grid in the box and only check atoms in neighboring grid
510 cells when constructing a new neighbor list every
511 :mdp:`nstlist` steps. In large systems grid search is much
512 faster than simple search.
514 .. mdp-value:: simple
516 Check every atom in the box when constructing a new neighbor
517 list every :mdp:`nstlist` steps (only with :mdp:`group`
524 Use periodic boundary conditions in all directions.
528 Use no periodic boundary conditions, ignore the box. To simulate
529 without cut-offs, set all cut-offs and :mdp:`nstlist` to 0. For
530 best performance without cut-offs on a single MPI rank, set
531 :mdp:`nstlist` to zero and :mdp:`ns-type` =simple.
535 Use periodic boundary conditions in x and y directions
536 only. This works only with :mdp:`ns-type` =grid and can be used
537 in combination with walls_. Without walls or with only one wall
538 the system size is infinite in the z direction. Therefore
539 pressure coupling or Ewald summation methods can not be
540 used. These disadvantages do not apply when two walls are used.
542 .. mdp:: periodic-molecules
546 molecules are finite, fast molecular PBC can be used
550 for systems with molecules that couple to themselves through the
551 periodic boundary conditions, this requires a slower PBC
552 algorithm and molecules are not made whole in the output
554 .. mdp:: verlet-buffer-tolerance
556 (0.005) \[kJ/mol/ps\]
558 Useful only with the :mdp:`Verlet` :mdp:`cutoff-scheme`. This sets
559 the maximum allowed error for pair interactions per particle caused
560 by the Verlet buffer, which indirectly sets :mdp:`rlist`. As both
561 :mdp:`nstlist` and the Verlet buffer size are fixed (for
562 performance reasons), particle pairs not in the pair list can
563 occasionally get within the cut-off distance during
564 :mdp:`nstlist` -1 steps. This causes very small jumps in the
565 energy. In a constant-temperature ensemble, these very small energy
566 jumps can be estimated for a given cut-off and :mdp:`rlist`. The
567 estimate assumes a homogeneous particle distribution, hence the
568 errors might be slightly underestimated for multi-phase
569 systems. (See the `reference manual`_ for details). For longer
570 pair-list life-time (:mdp:`nstlist` -1) * :mdp:`dt` the buffer is
571 overestimated, because the interactions between particles are
572 ignored. Combined with cancellation of errors, the actual drift of
573 the total energy is usually one to two orders of magnitude
574 smaller. Note that the generated buffer size takes into account
575 that the |Gromacs| pair-list setup leads to a reduction in the
576 drift by a factor 10, compared to a simple particle-pair based
577 list. Without dynamics (energy minimization etc.), the buffer is 5%
578 of the cut-off. For NVE simulations the initial temperature is
579 used, unless this is zero, in which case a buffer of 10% is
580 used. For NVE simulations the tolerance usually needs to be lowered
581 to achieve proper energy conservation on the nanosecond time
582 scale. To override the automated buffer setting, use
583 :mdp:`verlet-buffer-tolerance` =-1 and set :mdp:`rlist` manually.
588 Cut-off distance for the short-range neighbor list. With the
589 :mdp:`Verlet` :mdp:`cutoff-scheme`, this is by default set by the
590 :mdp:`verlet-buffer-tolerance` option and the value of
591 :mdp:`rlist` is ignored.
596 Cut-off distance for the long-range neighbor list. This parameter
597 is only relevant for a twin-range cut-off setup with switched
598 potentials. In that case a buffer region is required to account for
599 the size of charge groups. In all other cases this parameter is
600 automatically set to the longest cut-off distance.
608 .. mdp-value:: Cut-off
610 Twin range cut-offs with neighborlist cut-off :mdp:`rlist` and
611 Coulomb cut-off :mdp:`rcoulomb`, where :mdp:`rcoulomb` >=
616 Classical Ewald sum electrostatics. The real-space cut-off
617 :mdp:`rcoulomb` should be equal to :mdp:`rlist`. Use *e.g.*
618 :mdp:`rlist` =0.9, :mdp:`rcoulomb` =0.9. The highest magnitude
619 of wave vectors used in reciprocal space is controlled by
620 :mdp:`fourierspacing`. The relative accuracy of
621 direct/reciprocal space is controlled by :mdp:`ewald-rtol`.
623 NOTE: Ewald scales as O(N^3/2) and is thus extremely slow for
624 large systems. It is included mainly for reference - in most
625 cases PME will perform much better.
629 Fast smooth Particle-Mesh Ewald (SPME) electrostatics. Direct
630 space is similar to the Ewald sum, while the reciprocal part is
631 performed with FFTs. Grid dimensions are controlled with
632 :mdp:`fourierspacing` and the interpolation order with
633 :mdp:`pme-order`. With a grid spacing of 0.1 nm and cubic
634 interpolation the electrostatic forces have an accuracy of
635 2-3*10^-4. Since the error from the vdw-cutoff is larger than
636 this you might try 0.15 nm. When running in parallel the
637 interpolation parallelizes better than the FFT, so try
638 decreasing grid dimensions while increasing interpolation.
640 .. mdp-value:: P3M-AD
642 Particle-Particle Particle-Mesh algorithm with analytical
643 derivative for for long range electrostatic interactions. The
644 method and code is identical to SPME, except that the influence
645 function is optimized for the grid. This gives a slight increase
648 .. mdp-value:: Reaction-Field
650 Reaction field electrostatics with Coulomb cut-off
651 :mdp:`rcoulomb`, where :mdp:`rcoulomb` >= :mdp:`rlist`. The
652 dielectric constant beyond the cut-off is
653 :mdp:`epsilon-rf`. The dielectric constant can be set to
654 infinity by setting :mdp:`epsilon-rf` =0.
656 .. mdp-value:: Generalized-Reaction-Field
658 Generalized reaction field with Coulomb cut-off
659 :mdp:`rcoulomb`, where :mdp:`rcoulomb` >= :mdp:`rlist`. The
660 dielectric constant beyond the cut-off is
661 :mdp:`epsilon-rf`. The ionic strength is computed from the
662 number of charged (*i.e.* with non zero charge) charge
663 groups. The temperature for the GRF potential is set with
666 .. mdp-value:: Reaction-Field-zero
668 In |Gromacs|, normal reaction-field electrostatics with
669 :mdp:`cutoff-scheme` = :mdp:`group` leads to bad energy
670 conservation. :mdp:`Reaction-Field-zero` solves this by making
671 the potential zero beyond the cut-off. It can only be used with
672 an infinite dielectric constant (:mdp:`epsilon-rf` =0), because
673 only for that value the force vanishes at the
674 cut-off. :mdp:`rlist` should be 0.1 to 0.3 nm larger than
675 :mdp:`rcoulomb` to accommodate for the size of charge groups
676 and diffusion between neighbor list updates. This, and the fact
677 that table lookups are used instead of analytical functions make
678 :mdp:`Reaction-Field-zero` computationally more expensive than
679 normal reaction-field.
681 .. mdp-value:: Reaction-Field-nec
683 The same as :mdp-value:`coulombtype=Reaction-Field`, but
684 implemented as in |Gromacs| versions before 3.3. No
685 reaction-field correction is applied to excluded atom pairs and
686 self pairs. The 1-4 interactions are calculated using a
687 reaction-field. The missing correction due to the excluded pairs
688 that do not have a 1-4 interaction is up to a few percent of the
689 total electrostatic energy and causes a minor difference in the
690 forces and the pressure.
694 Analogous to :mdp-value:`vdwtype=Shift` for :mdp:`vdwtype`. You
695 might want to use :mdp:`Reaction-Field-zero` instead, which has
696 a similar potential shape, but has a physical interpretation and
697 has better energies due to the exclusion correction terms.
699 .. mdp-value:: Encad-Shift
701 The Coulomb potential is decreased over the whole range, using
702 the definition from the Encad simulation package.
704 .. mdp-value:: Switch
706 Analogous to :mdp-value:`vdwtype=Switch` for
707 :mdp:`vdwtype`. Switching the Coulomb potential can lead to
708 serious artifacts, advice: use :mdp:`Reaction-Field-zero`
713 :ref:`gmx mdrun` will now expect to find a file ``table.xvg``
714 with user-defined potential functions for repulsion, dispersion
715 and Coulomb. When pair interactions are present, :ref:`gmx
716 mdrun` also expects to find a file ``tablep.xvg`` for the pair
717 interactions. When the same interactions should be used for
718 non-bonded and pair interactions the user can specify the same
719 file name for both table files. These files should contain 7
720 columns: the ``x`` value, ``f(x)``, ``-f'(x)``, ``g(x)``,
721 ``-g'(x)``, ``h(x)``, ``-h'(x)``, where ``f(x)`` is the Coulomb
722 function, ``g(x)`` the dispersion function and ``h(x)`` the
723 repulsion function. When :mdp:`vdwtype` is not set to User the
724 values for ``g``, ``-g'``, ``h`` and ``-h'`` are ignored. For
725 the non-bonded interactions ``x`` values should run from 0 to
726 the largest cut-off distance + :mdp:`table-extension` and
727 should be uniformly spaced. For the pair interactions the table
728 length in the file will be used. The optimal spacing, which is
729 used for non-user tables, is ``0.002 nm`` when you run in mixed
730 precision or ``0.0005 nm`` when you run in double precision. The
731 function value at ``x=0`` is not important. More information is
732 in the printed manual.
734 .. mdp-value:: PME-Switch
736 A combination of PME and a switch function for the direct-space
737 part (see above). :mdp:`rcoulomb` is allowed to be smaller than
738 :mdp:`rlist`. This is mainly useful constant energy simulations
739 (note that using PME with :mdp:`cutoff-scheme` = :mdp:`Verlet`
740 will be more efficient).
742 .. mdp-value:: PME-User
744 A combination of PME and user tables (see
745 above). :mdp:`rcoulomb` is allowed to be smaller than
746 :mdp:`rlist`. The PME mesh contribution is subtracted from the
747 user table by :ref:`gmx mdrun`. Because of this subtraction the
748 user tables should contain about 10 decimal places.
750 .. mdp-value:: PME-User-Switch
752 A combination of PME-User and a switching function (see
753 above). The switching function is applied to final
754 particle-particle interaction, *i.e.* both to the user supplied
755 function and the PME Mesh correction part.
757 .. mdp:: coulomb-modifier
759 .. mdp-value:: Potential-shift-Verlet
761 Selects Potential-shift with the Verlet cutoff-scheme, as it is
762 (nearly) free; selects None with the group cutoff-scheme.
764 .. mdp-value:: Potential-shift
766 Shift the Coulomb potential by a constant such that it is zero
767 at the cut-off. This makes the potential the integral of the
768 force. Note that this does not affect the forces or the
773 Use an unmodified Coulomb potential. With the group scheme this
774 means no exact cut-off is used, energies and forces are
775 calculated for all pairs in the neighborlist.
777 .. mdp:: rcoulomb-switch
780 where to start switching the Coulomb potential, only relevant
781 when force or potential switching is used
786 distance for the Coulomb cut-off
791 The relative dielectric constant. A value of 0 means infinity.
796 The relative dielectric constant of the reaction field. This
797 is only used with reaction-field electrostatics. A value of 0
806 .. mdp-value:: Cut-off
808 Twin range cut-offs with neighbor list cut-off :mdp:`rlist` and
809 VdW cut-off :mdp:`rvdw`, where :mdp:`rvdw` >= :mdp:`rlist`.
813 Fast smooth Particle-mesh Ewald (SPME) for VdW interactions. The
814 grid dimensions are controlled with :mdp:`fourierspacing` in
815 the same way as for electrostatics, and the interpolation order
816 is controlled with :mdp:`pme-order`. The relative accuracy of
817 direct/reciprocal space is controlled by :mdp:`ewald-rtol-lj`,
818 and the specific combination rules that are to be used by the
819 reciprocal routine are set using :mdp:`lj-pme-comb-rule`.
823 This functionality is deprecated and replaced by
824 :mdp:`vdw-modifier` = Force-switch. The LJ (not Buckingham)
825 potential is decreased over the whole range and the forces decay
826 smoothly to zero between :mdp:`rvdw-switch` and
827 :mdp:`rvdw`. The neighbor search cut-off :mdp:`rlist` should
828 be 0.1 to 0.3 nm larger than :mdp:`rvdw` to accommodate for the
829 size of charge groups and diffusion between neighbor list
832 .. mdp-value:: Switch
834 This functionality is deprecated and replaced by
835 :mdp:`vdw-modifier` = Potential-switch. The LJ (not Buckingham)
836 potential is normal out to :mdp:`rvdw-switch`, after which it
837 is switched off to reach zero at :mdp:`rvdw`. Both the
838 potential and force functions are continuously smooth, but be
839 aware that all switch functions will give rise to a bulge
840 (increase) in the force (since we are switching the
841 potential). The neighbor search cut-off :mdp:`rlist` should be
842 0.1 to 0.3 nm larger than :mdp:`rvdw` to accommodate for the
843 size of charge groups and diffusion between neighbor list
846 .. mdp-value:: Encad-Shift
848 The LJ (not Buckingham) potential is decreased over the whole
849 range, using the definition from the Encad simulation package.
853 See user for :mdp:`coulombtype`. The function value at zero is
854 not important. When you want to use LJ correction, make sure
855 that :mdp:`rvdw` corresponds to the cut-off in the user-defined
856 function. When :mdp:`coulombtype` is not set to User the values
857 for the ``f`` and ``-f'`` columns are ignored.
859 .. mdp:: vdw-modifier
861 .. mdp-value:: Potential-shift-Verlet
863 Selects Potential-shift with the Verlet cutoff-scheme, as it is
864 (nearly) free; selects None with the group cutoff-scheme.
866 .. mdp-value:: Potential-shift
868 Shift the Van der Waals potential by a constant such that it is
869 zero at the cut-off. This makes the potential the integral of
870 the force. Note that this does not affect the forces or the
875 Use an unmodified Van der Waals potential. With the group scheme
876 this means no exact cut-off is used, energies and forces are
877 calculated for all pairs in the neighborlist.
879 .. mdp-value:: Force-switch
881 Smoothly switches the forces to zero between :mdp:`rvdw-switch`
882 and :mdp:`rvdw`. This shifts the potential shift over the whole
883 range and switches it to zero at the cut-off. Note that this is
884 more expensive to calculate than a plain cut-off and it is not
885 required for energy conservation, since Potential-shift
886 conserves energy just as well.
888 .. mdp-value:: Potential-switch
890 Smoothly switches the potential to zero between
891 :mdp:`rvdw-switch` and :mdp:`rvdw`. Note that this introduces
892 articifically large forces in the switching region and is much
893 more expensive to calculate. This option should only be used if
894 the force field you are using requires this.
900 where to start switching the LJ force and possibly the potential,
901 only relevant when force or potential switching is used
906 distance for the LJ or Buckingham cut-off
912 don't apply any correction
914 .. mdp-value:: EnerPres
916 apply long range dispersion corrections for Energy and Pressure
920 apply long range dispersion corrections for Energy only
926 .. mdp:: table-extension
929 Extension of the non-bonded potential lookup tables beyond the
930 largest cut-off distance. The value should be large enough to
931 account for charge group sizes and the diffusion between
932 neighbor-list updates. Without user defined potential the same
933 table length is used for the lookup tables for the 1-4
934 interactions, which are always tabulated irrespective of the use of
935 tables for the non-bonded interactions. The value of
936 :mdp:`table-extension` in no way affects the values of
937 :mdp:`rlist`, :mdp:`rcoulomb`, or :mdp:`rvdw`.
939 .. mdp:: energygrp-table
941 When user tables are used for electrostatics and/or VdW, here one
942 can give pairs of energy groups for which seperate user tables
943 should be used. The two energy groups will be appended to the table
944 file name, in order of their definition in :mdp:`energygrps`,
945 seperated by underscores. For example, if ``energygrps = Na Cl
946 Sol`` and ``energygrp-table = Na Na Na Cl``, :ref:`gmx mdrun` will
947 read ``table_Na_Na.xvg`` and ``table_Na_Cl.xvg`` in addition to the
948 normal ``table.xvg`` which will be used for all other energy group
955 .. mdp:: fourierspacing
958 For ordinary Ewald, the ratio of the box dimensions and the spacing
959 determines a lower bound for the number of wave vectors to use in
960 each (signed) direction. For PME and P3M, that ratio determines a
961 lower bound for the number of Fourier-space grid points that will
962 be used along that axis. In all cases, the number for each
963 direction can be overridden by entering a non-zero value for that
964 :mdp:`fourier-nx` direction. For optimizing the relative load of
965 the particle-particle interactions and the mesh part of PME, it is
966 useful to know that the accuracy of the electrostatics remains
967 nearly constant when the Coulomb cut-off and the PME grid spacing
968 are scaled by the same factor.
975 Highest magnitude of wave vectors in reciprocal space when using Ewald.
976 Grid size when using PME or P3M. These values override
977 :mdp:`fourierspacing` per direction. The best choice is powers of
978 2, 3, 5 and 7. Avoid large primes.
983 Interpolation order for PME. 4 equals cubic interpolation. You
984 might try 6/8/10 when running in parallel and simultaneously
985 decrease grid dimension.
990 The relative strength of the Ewald-shifted direct potential at
991 :mdp:`rcoulomb` is given by :mdp:`ewald-rtol`. Decreasing this
992 will give a more accurate direct sum, but then you need more wave
993 vectors for the reciprocal sum.
995 .. mdp:: ewald-rtol-lj
998 When doing PME for VdW-interactions, :mdp:`ewald-rtol-lj` is used
999 to control the relative strength of the dispersion potential at
1000 :mdp:`rvdw` in the same way as :mdp:`ewald-rtol` controls the
1001 electrostatic potential.
1003 .. mdp:: lj-pme-comb-rule
1006 The combination rules used to combine VdW-parameters in the
1007 reciprocal part of LJ-PME. Geometric rules are much faster than
1008 Lorentz-Berthelot and usually the recommended choice, even when the
1009 rest of the force field uses the Lorentz-Berthelot rules.
1011 .. mdp-value:: Geometric
1013 Apply geometric combination rules
1015 .. mdp-value:: Lorentz-Berthelot
1017 Apply Lorentz-Berthelot combination rules
1019 .. mdp:: ewald-geometry
1023 The Ewald sum is performed in all three dimensions.
1027 The reciprocal sum is still performed in 3D, but a force and
1028 potential correction applied in the `z` dimension to produce a
1029 pseudo-2D summation. If your system has a slab geometry in the
1030 `x-y` plane you can try to increase the `z`-dimension of the box
1031 (a box height of 3 times the slab height is usually ok) and use
1034 .. mdp:: epsilon-surface
1037 This controls the dipole correction to the Ewald summation in
1038 3D. The default value of zero means it is turned off. Turn it on by
1039 setting it to the value of the relative permittivity of the
1040 imaginary surface around your infinite system. Be careful - you
1041 shouldn't use this if you have free mobile charges in your
1042 system. This value does not affect the slab 3DC variant of the long
1046 Temperature coupling
1047 ^^^^^^^^^^^^^^^^^^^^
1053 No temperature coupling.
1055 .. mdp-value:: berendsen
1057 Temperature coupling with a Berendsen-thermostat to a bath with
1058 temperature :mdp:`ref-t`, with time constant
1059 :mdp:`tau-t`. Several groups can be coupled separately, these
1060 are specified in the :mdp:`tc-grps` field separated by spaces.
1062 .. mdp-value:: nose-hoover
1064 Temperature coupling using a Nose-Hoover extended ensemble. The
1065 reference temperature and coupling groups are selected as above,
1066 but in this case :mdp:`tau-t` controls the period of the
1067 temperature fluctuations at equilibrium, which is slightly
1068 different from a relaxation time. For NVT simulations the
1069 conserved energy quantity is written to energy and log file.
1071 .. mdp-value:: andersen
1073 Temperature coupling by randomizing a fraction of the particles
1074 at each timestep. Reference temperature and coupling groups are
1075 selected as above. :mdp:`tau-t` is the average time between
1076 randomization of each molecule. Inhibits particle dynamics
1077 somewhat, but little or no ergodicity issues. Currently only
1078 implemented with velocity Verlet, and not implemented with
1081 .. mdp-value:: andersen-massive
1083 Temperature coupling by randomizing all particles at infrequent
1084 timesteps. Reference temperature and coupling groups are
1085 selected as above. :mdp:`tau-t` is the time between
1086 randomization of all molecules. Inhibits particle dynamics
1087 somewhat, but little or no ergodicity issues. Currently only
1088 implemented with velocity Verlet.
1090 .. mdp-value:: v-rescale
1092 Temperature coupling using velocity rescaling with a stochastic
1093 term (JCP 126, 014101). This thermostat is similar to Berendsen
1094 coupling, with the same scaling using :mdp:`tau-t`, but the
1095 stochastic term ensures that a proper canonical ensemble is
1096 generated. The random seed is set with :mdp:`ld-seed`. This
1097 thermostat works correctly even for :mdp:`tau-t` =0. For NVT
1098 simulations the conserved energy quantity is written to the
1099 energy and log file.
1104 The frequency for coupling the temperature. The default value of -1
1105 sets :mdp:`nsttcouple` equal to :mdp:`nstlist`, unless
1106 :mdp:`nstlist` <=0, then a value of 10 is used. For velocity
1107 Verlet integrators :mdp:`nsttcouple` is set to 1.
1109 .. mdp:: nh-chain-length
1112 The number of chained Nose-Hoover thermostats for velocity Verlet
1113 integrators, the leap-frog :mdp-value:`integrator=md` integrator
1114 only supports 1. Data for the NH chain variables is not printed to
1115 the :ref:`edr` file, but can be using the ``GMX_NOSEHOOVER_CHAINS``
1116 environment variable
1120 groups to couple to separate temperature baths
1125 time constant for coupling (one for each group in
1126 :mdp:`tc-grps`), -1 means no temperature coupling
1131 reference temperature for coupling (one for each group in
1142 No pressure coupling. This means a fixed box size.
1144 .. mdp-value:: Berendsen
1146 Exponential relaxation pressure coupling with time constant
1147 :mdp:`tau-p`. The box is scaled every timestep. It has been
1148 argued that this does not yield a correct thermodynamic
1149 ensemble, but it is the most efficient way to scale a box at the
1152 .. mdp-value:: Parrinello-Rahman
1154 Extended-ensemble pressure coupling where the box vectors are
1155 subject to an equation of motion. The equation of motion for the
1156 atoms is coupled to this. No instantaneous scaling takes
1157 place. As for Nose-Hoover temperature coupling the time constant
1158 :mdp:`tau-p` is the period of pressure fluctuations at
1159 equilibrium. This is probably a better method when you want to
1160 apply pressure scaling during data collection, but beware that
1161 you can get very large oscillations if you are starting from a
1162 different pressure. For simulations where the exact fluctation
1163 of the NPT ensemble are important, or if the pressure coupling
1164 time is very short it may not be appropriate, as the previous
1165 time step pressure is used in some steps of the |Gromacs|
1166 implementation for the current time step pressure.
1170 Martyna-Tuckerman-Tobias-Klein implementation, only useable with
1171 :mdp-value:`md-vv` or :mdp-value:`md-vv-avek`, very similar to
1172 Parrinello-Rahman. As for Nose-Hoover temperature coupling the
1173 time constant :mdp:`tau-p` is the period of pressure
1174 fluctuations at equilibrium. This is probably a better method
1175 when you want to apply pressure scaling during data collection,
1176 but beware that you can get very large oscillations if you are
1177 starting from a different pressure. Currently (as of version
1178 5.1), it only supports isotropic scaling, and only works without
1183 .. mdp-value:: isotropic
1185 Isotropic pressure coupling with time constant
1186 :mdp:`tau-p`. The compressibility and reference pressure are
1187 set with :mdp:`compressibility` and :mdp:`ref-p`, one value is
1190 .. mdp-value:: semiisotropic
1192 Pressure coupling which is isotropic in the ``x`` and ``y``
1193 direction, but different in the ``z`` direction. This can be
1194 useful for membrane simulations. 2 values are needed for ``x/y``
1195 and ``z`` directions respectively.
1197 .. mdp-value:: anisotropic
1199 Same as before, but 6 values are needed for ``xx``, ``yy``, ``zz``,
1200 ``xy/yx``, ``xz/zx`` and ``yz/zy`` components,
1201 respectively. When the off-diagonal compressibilities are set to
1202 zero, a rectangular box will stay rectangular. Beware that
1203 anisotropic scaling can lead to extreme deformation of the
1206 .. mdp-value:: surface-tension
1208 Surface tension coupling for surfaces parallel to the
1209 xy-plane. Uses normal pressure coupling for the `z`-direction,
1210 while the surface tension is coupled to the `x/y` dimensions of
1211 the box. The first :mdp:`ref-p` value is the reference surface
1212 tension times the number of surfaces ``bar nm``, the second
1213 value is the reference `z`-pressure ``bar``. The two
1214 :mdp:`compressibility` values are the compressibility in the
1215 `x/y` and `z` direction respectively. The value for the
1216 `z`-compressibility should be reasonably accurate since it
1217 influences the convergence of the surface-tension, it can also
1218 be set to zero to have a box with constant height.
1223 The frequency for coupling the pressure. The default value of -1
1224 sets :mdp:`nstpcouple` equal to :mdp:`nstlist`, unless
1225 :mdp:`nstlist` <=0, then a value of 10 is used. For velocity
1226 Verlet integrators :mdp:`nstpcouple` is set to 1.
1231 time constant for coupling
1233 .. mdp:: compressibility
1236 compressibility (NOTE: this is now really in bar-1) For water at 1
1237 atm and 300 K the compressibility is 4.5e-5 bar^-1.
1242 reference pressure for coupling
1244 .. mdp:: refcoord-scaling
1248 The reference coordinates for position restraints are not
1249 modified. Note that with this option the virial and pressure
1250 will depend on the absolute positions of the reference
1255 The reference coordinates are scaled with the scaling matrix of
1256 the pressure coupling.
1260 Scale the center of mass of the reference coordinates with the
1261 scaling matrix of the pressure coupling. The vectors of each
1262 reference coordinate to the center of mass are not scaled. Only
1263 one COM is used, even when there are multiple molecules with
1264 position restraints. For calculating the COM of the reference
1265 coordinates in the starting configuration, periodic boundary
1266 conditions are not taken into account.
1272 Simulated annealing is controlled separately for each temperature
1273 group in |Gromacs|. The reference temperature is a piecewise linear
1274 function, but you can use an arbitrary number of points for each
1275 group, and choose either a single sequence or a periodic behaviour for
1276 each group. The actual annealing is performed by dynamically changing
1277 the reference temperature used in the thermostat algorithm selected,
1278 so remember that the system will usually not instantaneously reach the
1279 reference temperature!
1283 Type of annealing for each temperature group
1287 No simulated annealing - just couple to reference temperature value.
1289 .. mdp-value:: single
1291 A single sequence of annealing points. If your simulation is
1292 longer than the time of the last point, the temperature will be
1293 coupled to this constant value after the annealing sequence has
1294 reached the last time point.
1296 .. mdp-value:: periodic
1298 The annealing will start over at the first reference point once
1299 the last reference time is reached. This is repeated until the
1302 .. mdp:: annealing-npoints
1304 A list with the number of annealing reference/control points used
1305 for each temperature group. Use 0 for groups that are not
1306 annealed. The number of entries should equal the number of
1309 .. mdp:: annealing-time
1311 List of times at the annealing reference/control points for each
1312 group. If you are using periodic annealing, the times will be used
1313 modulo the last value, *i.e.* if the values are 0, 5, 10, and 15,
1314 the coupling will restart at the 0ps value after 15ps, 30ps, 45ps,
1315 etc. The number of entries should equal the sum of the numbers
1316 given in :mdp:`annealing-npoints`.
1318 .. mdp:: annealing-temp
1320 List of temperatures at the annealing reference/control points for
1321 each group. The number of entries should equal the sum of the
1322 numbers given in :mdp:`annealing-npoints`.
1324 Confused? OK, let's use an example. Assume you have two temperature
1325 groups, set the group selections to ``annealing = single periodic``,
1326 the number of points of each group to ``annealing-npoints = 3 4``, the
1327 times to ``annealing-time = 0 3 6 0 2 4 6`` and finally temperatures
1328 to ``annealing-temp = 298 280 270 298 320 320 298``. The first group
1329 will be coupled to 298K at 0ps, but the reference temperature will
1330 drop linearly to reach 280K at 3ps, and then linearly between 280K and
1331 270K from 3ps to 6ps. After this is stays constant, at 270K. The
1332 second group is coupled to 298K at 0ps, it increases linearly to 320K
1333 at 2ps, where it stays constant until 4ps. Between 4ps and 6ps it
1334 decreases to 298K, and then it starts over with the same pattern
1335 again, *i.e.* rising linearly from 298K to 320K between 6ps and
1336 8ps. Check the summary printed by :ref:`gmx grompp` if you are unsure!
1346 Do not generate velocities. The velocities are set to zero
1347 when there are no velocities in the input structure file.
1351 Generate velocities in :ref:`gmx grompp` according to a
1352 Maxwell distribution at temperature :mdp:`gen-temp`, with
1353 random seed :mdp:`gen-seed`. This is only meaningful with
1354 integrator :mdp-value:`integrator=md`.
1359 temperature for Maxwell distribution
1364 used to initialize random generator for random velocities,
1365 when :mdp:`gen-seed` is set to -1, a pseudo random seed is
1372 .. mdp:: constraints
1376 No constraints except for those defined explicitly in the
1377 topology, *i.e.* bonds are represented by a harmonic (or other)
1378 potential or a Morse potential (depending on the setting of
1379 :mdp:`morse`) and angles by a harmonic (or other) potential.
1381 .. mdp-value:: h-bonds
1383 Convert the bonds with H-atoms to constraints.
1385 .. mdp-value:: all-bonds
1387 Convert all bonds to constraints.
1389 .. mdp-value:: h-angles
1391 Convert all bonds and additionally the angles that involve
1392 H-atoms to bond-constraints.
1394 .. mdp-value:: all-angles
1396 Convert all bonds and angles to bond-constraints.
1398 .. mdp:: constraint-algorithm
1400 .. mdp-value:: LINCS
1402 LINear Constraint Solver. With domain decomposition the parallel
1403 version P-LINCS is used. The accuracy in set with
1404 :mdp:`lincs-order`, which sets the number of matrices in the
1405 expansion for the matrix inversion. After the matrix inversion
1406 correction the algorithm does an iterative correction to
1407 compensate for lengthening due to rotation. The number of such
1408 iterations can be controlled with :mdp:`lincs-iter`. The root
1409 mean square relative constraint deviation is printed to the log
1410 file every :mdp:`nstlog` steps. If a bond rotates more than
1411 :mdp:`lincs-warnangle` in one step, a warning will be printed
1412 both to the log file and to ``stderr``. LINCS should not be used
1413 with coupled angle constraints.
1415 .. mdp-value:: SHAKE
1417 SHAKE is slightly slower and less stable than LINCS, but does
1418 work with angle constraints. The relative tolerance is set with
1419 :mdp:`shake-tol`, 0.0001 is a good value for "normal" MD. SHAKE
1420 does not support constraints between atoms on different nodes,
1421 thus it can not be used with domain decompositon when inter
1422 charge-group constraints are present. SHAKE can not be used with
1423 energy minimization.
1425 .. mdp:: continuation
1427 This option was formerly known as unconstrained-start.
1431 apply constraints to the start configuration and reset shells
1435 do not apply constraints to the start configuration and do not
1436 reset shells, useful for exact coninuation and reruns
1441 relative tolerance for SHAKE
1443 .. mdp:: lincs-order
1446 Highest order in the expansion of the constraint coupling
1447 matrix. When constraints form triangles, an additional expansion of
1448 the same order is applied on top of the normal expansion only for
1449 the couplings within such triangles. For "normal" MD simulations an
1450 order of 4 usually suffices, 6 is needed for large time-steps with
1451 virtual sites or BD. For accurate energy minimization an order of 8
1452 or more might be required. With domain decomposition, the cell size
1453 is limited by the distance spanned by :mdp:`lincs-order` +1
1454 constraints. When one wants to scale further than this limit, one
1455 can decrease :mdp:`lincs-order` and increase :mdp:`lincs-iter`,
1456 since the accuracy does not deteriorate when (1+ :mdp:`lincs-iter`
1457 )* :mdp:`lincs-order` remains constant.
1462 Number of iterations to correct for rotational lengthening in
1463 LINCS. For normal runs a single step is sufficient, but for NVE
1464 runs where you want to conserve energy accurately or for accurate
1465 energy minimization you might want to increase it to 2.
1467 .. mdp:: lincs-warnangle
1470 maximum angle that a bond can rotate before LINCS will complain
1476 bonds are represented by a harmonic potential
1480 bonds are represented by a Morse potential
1483 Energy group exclusions
1484 ^^^^^^^^^^^^^^^^^^^^^^^
1486 .. mdp:: energygrp-excl:
1488 Pairs of energy groups for which all non-bonded interactions are
1489 excluded. An example: if you have two energy groups ``Protein`` and
1490 ``SOL``, specifying ``energygrp-excl = Protein Protein SOL SOL``
1491 would give only the non-bonded interactions between the protein and
1492 the solvent. This is especially useful for speeding up energy
1493 calculations with ``mdrun -rerun`` and for excluding interactions
1494 within frozen groups.
1503 When set to 1 there is a wall at ``z=0``, when set to 2 there is
1504 also a wall at ``z=z-box``. Walls can only be used with :mdp:`pbc`
1505 ``=xy``. When set to 2 pressure coupling and Ewald summation can be
1506 used (it is usually best to use semiisotropic pressure coupling
1507 with the ``x/y`` compressibility set to 0, as otherwise the surface
1508 area will change). Walls interact wit the rest of the system
1509 through an optional :mdp:`wall-atomtype`. Energy groups ``wall0``
1510 and ``wall1`` (for :mdp:`nwall` =2) are added automatically to
1511 monitor the interaction of energy groups with each wall. The center
1512 of mass motion removal will be turned off in the ``z``-direction.
1514 .. mdp:: wall-atomtype
1516 the atom type name in the force field for each wall. By (for
1517 example) defining a special wall atom type in the topology with its
1518 own combination rules, this allows for independent tuning of the
1519 interaction of each atomtype with the walls.
1525 LJ integrated over the volume behind the wall: 9-3 potential
1529 LJ integrated over the wall surface: 10-4 potential
1533 direct LJ potential with the ``z`` distance from the wall
1537 user defined potentials indexed with the ``z`` distance from the
1538 wall, the tables are read analogously to the
1539 :mdp:`energygrp-table` option, where the first name is for a
1540 "normal" energy group and the second name is ``wall0`` or
1541 ``wall1``, only the dispersion and repulsion columns are used
1543 .. mdp:: wall-r-linpot
1546 Below this distance from the wall the potential is continued
1547 linearly and thus the force is constant. Setting this option to a
1548 postive value is especially useful for equilibration when some
1549 atoms are beyond a wall. When the value is <=0 (<0 for
1550 :mdp:`wall-type` =table), a fatal error is generated when atoms
1553 .. mdp:: wall-density
1556 the number density of the atoms for each wall for wall types 9-3
1559 .. mdp:: wall-ewald-zfac
1562 The scaling factor for the third box vector for Ewald summation
1563 only, the minimum is 2. Ewald summation can only be used with
1564 :mdp:`nwall` =2, where one should use :mdp:`ewald-geometry`
1565 ``=3dc``. The empty layer in the box serves to decrease the
1566 unphysical Coulomb interaction between periodic images.
1572 Note that where pulling coordinate are applicable, there can be more
1573 than one (set with :mdp:`pull-ncoords`) and multiple related :ref:`mdp`
1574 variables will exist accordingly. Documentation references to things
1575 like :mdp:`pull-coord1-vec` should be understood to apply to to the
1576 applicable pulling coordinate.
1582 No center of mass pulling. All the following pull options will
1583 be ignored (and if present in the :ref:`mdp` file, they unfortunately
1588 Center of mass pulling will be applied on 1 or more groups using
1589 1 or more pull coordinates.
1591 .. mdp:: pull-cylinder-r
1594 the radius of the cylinder for
1595 :mdp:`pull-coord1-geometry` = :mdp-value:`cylinder`
1597 .. mdp:: pull-constr-tol
1600 the relative constraint tolerance for constraint pulling
1602 .. mdp:: pull-print-com1
1606 do not print the COM of the first group in each pull coordinate
1610 print the COM of the first group in each pull coordinate
1612 .. mdp:: pull-print-com2
1616 do not print the COM of the second group in each pull coordinate
1620 print the COM of the second group in each pull coordinate
1622 .. mdp:: pull-print-ref-value
1626 do not print the reference value for each pull coordinate
1630 print the reference value for each pull coordinate
1632 .. mdp:: pull-print-components
1636 only print the distance for each pull coordinate
1640 print the distance and Cartesian components selected in
1641 :mdp:`pull-coord1-dim`
1643 .. mdp:: pull-nstxout
1646 frequency for writing out the COMs of all the pull group (0 is
1649 .. mdp:: pull-nstfout
1652 frequency for writing out the force of all the pulled group
1656 .. mdp:: pull-ngroups
1659 The number of pull groups, not including the absolute reference
1660 group, when used. Pull groups can be reused in multiple pull
1661 coordinates. Below only the pull options for group 1 are given,
1662 further groups simply increase the group index number.
1664 .. mdp:: pull-ncoords
1667 The number of pull coordinates. Below only the pull options for
1668 coordinate 1 are given, further coordinates simply increase the
1669 coordinate index number.
1671 .. mdp:: pull-group1-name
1673 The name of the pull group, is looked up in the index file or in
1674 the default groups to obtain the atoms involved.
1676 .. mdp:: pull-group1-weights
1678 Optional relative weights which are multiplied with the masses of
1679 the atoms to give the total weight for the COM. The number should
1680 be 0, meaning all 1, or the number of atoms in the pull group.
1682 .. mdp:: pull-group1-pbcatom
1685 The reference atom for the treatment of periodic boundary
1686 conditions inside the group (this has no effect on the treatment of
1687 the pbc between groups). This option is only important when the
1688 diameter of the pull group is larger than half the shortest box
1689 vector. For determining the COM, all atoms in the group are put at
1690 their periodic image which is closest to
1691 :mdp:`pull-group1-pbcatom`. A value of 0 means that the middle
1692 atom (number wise) is used. This parameter is not used with
1693 :mdp:`pull-geometry` cylinder. A value of -1 turns on cosine
1694 weighting, which is useful for a group of molecules in a periodic
1695 system, *e.g.* a water slab (see Engin et al. J. Chem. Phys. B
1698 .. mdp:: pull-coord1-groups
1700 The two groups indices should be given on which this pull
1701 coordinate will operate. The first index can be 0, in which case an
1702 absolute reference of :mdp:`pull-coord1-origin` is used. With an
1703 absolute reference the system is no longer translation invariant
1704 and one should think about what to do with the center of mass
1707 .. mdp:: pull-coord1-type:
1709 .. mdp-value:: umbrella
1711 Center of mass pulling using an umbrella potential between the
1712 reference group and one or more groups.
1714 .. mdp-value:: constraint
1716 Center of mass pulling using a constraint between the reference
1717 group and one or more groups. The setup is identical to the
1718 option umbrella, except for the fact that a rigid constraint is
1719 applied instead of a harmonic potential.
1721 .. mdp-value:: constant-force
1723 Center of mass pulling using a linear potential and therefore a
1724 constant force. For this option there is no reference position
1725 and therefore the parameters :mdp:`pull-coord1-init` and
1726 :mdp:`pull-coord1-rate` are not used.
1728 .. mdp-value:: flat-bottom
1730 At distances beyond :mdp:`pull-coord1-init` a harmonic potential
1731 is applied, otherwise no potential is applied.
1733 .. mdp:: pull-coord1-geometry
1735 .. mdp-value:: distance
1737 Pull along the vector connecting the two groups. Components can
1738 be selected with :mdp:`pull-coord1-dim`.
1740 .. mdp-value:: direction
1742 Pull in the direction of :mdp:`pull-coord1-vec`.
1744 .. mdp-value:: direction-periodic
1746 As :mdp-value:`direction`, but allows the distance to be larger
1747 than half the box size. With this geometry the box should not be
1748 dynamic (*e.g.* no pressure scaling) in the pull dimensions and
1749 the pull force is not added to virial.
1751 .. mdp-value:: cylinder
1753 Designed for pulling with respect to a layer where the reference
1754 COM is given by a local cylindrical part of the reference group.
1755 The pulling is in the direction of :mdp:`pull-coord1-vec`. From
1756 the first of the two groups in :mdp:`pull-coord1-groups` a
1757 cylinder is selected around the axis going through the COM of
1758 the second group with direction :mdp:`pull-coord1-vec` with
1759 radius :mdp:`pull-cylinder-r`. Weights of the atoms decrease
1760 continously to zero as the radial distance goes from 0 to
1761 :mdp:`pull-cylinder-r` (mass weighting is also used). The radial
1762 dependence gives rise to radial forces on both pull groups.
1763 Note that the radius should be smaller than half the box size.
1764 For tilted cylinders they should be even smaller than half the
1765 box size since the distance of an atom in the reference group
1766 from the COM of the pull group has both a radial and an axial
1767 component. This geometry is not supported with constraint
1770 .. mdp:: pull-coord1-dim
1773 Selects the dimensions that this pull coordinate acts on and that
1774 are printed to the output files when
1775 :mdp:`pull-print-components` = :mdp-value:`yes`. With
1776 :mdp:`pull-coord1-geometry` = :mdp-value:`distance`, only Cartesian
1777 components set to Y contribute to the distance. Thus setting this
1778 to Y Y N results in a distance in the x/y plane. With other
1779 geometries all dimensions with non-zero entries in
1780 :mdp:`pull-coord1-vec` should be set to Y, the values for other
1781 dimensions only affect the output.
1783 .. mdp:: pull-coord1-origin
1786 The pull reference position for use with an absolute reference.
1788 .. mdp:: pull-coord1-vec
1791 The pull direction. :ref:`gmx grompp` normalizes the vector.
1793 .. mdp:: pull-coord1-start
1797 do not modify :mdp:`pull-coord1-init`
1801 add the COM distance of the starting conformation to
1802 :mdp:`pull-coord1-init`
1804 .. mdp:: pull-coord1-init
1807 The reference distance at t=0.
1809 .. mdp:: pull-coord1-rate
1812 The rate of change of the reference position.
1814 .. mdp:: pull-coord1-k
1816 (0) \[kJ mol-1 nm-2\] / \[kJ mol-1 nm-1\]
1817 The force constant. For umbrella pulling this is the harmonic force
1818 constant in kJ mol-1 nm-2. For constant force pulling this is the
1819 force constant of the linear potential, and thus the negative (!)
1820 of the constant force in kJ mol-1 nm-1.
1822 .. mdp:: pull-coord1-kB
1824 (pull-k1) \[kJ mol-1 nm-2\] / \[kJ mol-1 nm-1\]
1825 As :mdp:`pull-coord1-k`, but for state B. This is only used when
1826 :mdp:`free-energy` is turned on. The force constant is then (1 -
1827 lambda) * :mdp:`pull-coord1-k` + lambda * :mdp:`pull-coord1-kB`.
1837 ignore distance restraint information in topology file
1839 .. mdp-value:: simple
1841 simple (per-molecule) distance restraints.
1843 .. mdp-value:: ensemble
1845 distance restraints over an ensemble of molecules in one
1846 simulation box. Normally, one would perform ensemble averaging
1847 over multiple subsystems, each in a separate box, using ``mdrun
1848 -multi``. Supply ``topol0.tpr``, ``topol1.tpr``, ... with
1849 different coordinates and/or velocities. The environment
1850 variable ``GMX_DISRE_ENSEMBLE_SIZE`` sets the number of systems
1851 within each ensemble (usually equal to the ``mdrun -multi``
1854 .. mdp:: disre-weighting
1856 .. mdp-value:: equal
1858 divide the restraint force equally over all atom pairs in the
1861 .. mdp-value:: conservative
1863 the forces are the derivative of the restraint potential, this
1864 results in an weighting of the atom pairs to the reciprocal
1865 seventh power of the displacement. The forces are conservative
1866 when :mdp:`disre-tau` is zero.
1868 .. mdp:: disre-mixed
1872 the violation used in the calculation of the restraint force is
1873 the time-averaged violation
1877 the violation used in the calculation of the restraint force is
1878 the square root of the product of the time-averaged violation
1879 and the instantaneous violation
1883 (1000) \[kJ mol-1 nm-2\]
1884 force constant for distance restraints, which is multiplied by a
1885 (possibly) different factor for each restraint given in the `fac`
1886 column of the interaction in the topology file.
1891 time constant for distance restraints running average. A value of
1892 zero turns off time averaging.
1894 .. mdp:: nstdisreout
1897 period between steps when the running time-averaged and
1898 instantaneous distances of all atom pairs involved in restraints
1899 are written to the energy file (can make the energy file very
1906 ignore orientation restraint information in topology file
1910 use orientation restraints, ensemble averaging can be performed
1916 force constant for orientation restraints, which is multiplied by a
1917 (possibly) different weight factor for each restraint, can be set
1918 to zero to obtain the orientations from a free simulation
1923 time constant for orientation restraints running average. A value
1924 of zero turns off time averaging.
1926 .. mdp:: orire-fitgrp
1928 fit group for orientation restraining. This group of atoms is used
1929 to determine the rotation **R** of the system with respect to the
1930 reference orientation. The reference orientation is the starting
1931 conformation of the first subsystem. For a protein, backbone is a
1934 .. mdp:: nstorireout
1937 period between steps when the running time-averaged and
1938 instantaneous orientations for all restraints, and the molecular
1939 order tensor are written to the energy file (can make the energy
1943 Free energy calculations
1944 ^^^^^^^^^^^^^^^^^^^^^^^^
1946 .. mdp:: free-energy
1950 Only use topology A.
1954 Interpolate between topology A (lambda=0) to topology B
1955 (lambda=1) and write the derivative of the Hamiltonian with
1956 respect to lambda (as specified with :mdp:`dhdl-derivatives`),
1957 or the Hamiltonian differences with respect to other lambda
1958 values (as specified with foreign lambda) to the energy file
1959 and/or to ``dhdl.xvg``, where they can be processed by, for
1960 example :ref:`gmx bar`. The potentials, bond-lengths and angles
1961 are interpolated linearly as described in the manual. When
1962 :mdp:`sc-alpha` is larger than zero, soft-core potentials are
1963 used for the LJ and Coulomb interactions.
1967 Turns on expanded ensemble simulation, where the alchemical state
1968 becomes a dynamic variable, allowing jumping between different
1969 Hamiltonians. See the expanded ensemble options for controlling how
1970 expanded ensemble simulations are performed. The different
1971 Hamiltonians used in expanded ensemble simulations are defined by
1972 the other free energy options.
1974 .. mdp:: init-lambda
1977 starting value for lambda (float). Generally, this should only be
1978 used with slow growth (*i.e.* nonzero :mdp:`delta-lambda`). In
1979 other cases, :mdp:`init-lambda-state` should be specified
1980 instead. Must be greater than or equal to 0.
1982 .. mdp:: delta-lambda
1985 increment per time step for lambda
1987 .. mdp:: init-lambda-state
1990 starting value for the lambda state (integer). Specifies which
1991 columm of the lambda vector (:mdp:`coul-lambdas`,
1992 :mdp:`vdw-lambdas`, :mdp:`bonded-lambdas`,
1993 :mdp:`restraint-lambdas`, :mdp:`mass-lambdas`,
1994 :mdp:`temperature-lambdas`, :mdp:`fep-lambdas`) should be
1995 used. This is a zero-based index: :mdp:`init-lambda-state` 0 means
1996 the first column, and so on.
1998 .. mdp:: fep-lambdas
2001 Zero, one or more lambda values for which Delta H values will be
2002 determined and written to dhdl.xvg every :mdp:`nstdhdl`
2003 steps. Values must be between 0 and 1. Free energy differences
2004 between different lambda values can then be determined with
2005 :ref:`gmx bar`. :mdp:`fep-lambdas` is different from the
2006 other -lambdas keywords because all components of the lambda vector
2007 that are not specified will use :mdp:`fep-lambdas` (including
2008 :mdp:`restraint-lambdas` and therefore the pull code restraints).
2010 .. mdp:: coul-lambdas
2013 Zero, one or more lambda values for which Delta H values will be
2014 determined and written to dhdl.xvg every :mdp:`nstdhdl`
2015 steps. Values must be between 0 and 1. Only the electrostatic
2016 interactions are controlled with this component of the lambda
2017 vector (and only if the lambda=0 and lambda=1 states have differing
2018 electrostatic interactions).
2020 .. mdp:: vdw-lambdas
2023 Zero, one or more lambda values for which Delta H values will be
2024 determined and written to dhdl.xvg every :mdp:`nstdhdl`
2025 steps. Values must be between 0 and 1. Only the van der Waals
2026 interactions are controlled with this component of the lambda
2029 .. mdp:: bonded-lambdas
2032 Zero, one or more lambda values for which Delta H values will be
2033 determined and written to dhdl.xvg every :mdp:`nstdhdl`
2034 steps. Values must be between 0 and 1. Only the bonded interactions
2035 are controlled with this component of the lambda vector.
2037 .. mdp:: restraint-lambdas
2040 Zero, one or more lambda values for which Delta H values will be
2041 determined and written to dhdl.xvg every :mdp:`nstdhdl`
2042 steps. Values must be between 0 and 1. Only the restraint
2043 interactions: dihedral restraints, and the pull code restraints are
2044 controlled with this component of the lambda vector.
2046 .. mdp:: mass-lambdas
2049 Zero, one or more lambda values for which Delta H values will be
2050 determined and written to dhdl.xvg every :mdp:`nstdhdl`
2051 steps. Values must be between 0 and 1. Only the particle masses are
2052 controlled with this component of the lambda vector.
2054 .. mdp:: temperature-lambdas
2057 Zero, one or more lambda values for which Delta H values will be
2058 determined and written to dhdl.xvg every :mdp:`nstdhdl`
2059 steps. Values must be between 0 and 1. Only the temperatures
2060 controlled with this component of the lambda vector. Note that
2061 these lambdas should not be used for replica exchange, only for
2062 simulated tempering.
2064 .. mdp:: calc-lambda-neighbors
2067 Controls the number of lambda values for which Delta H values will
2068 be calculated and written out, if :mdp:`init-lambda-state` has
2069 been set. A positive value will limit the number of lambda points
2070 calculated to only the nth neighbors of :mdp:`init-lambda-state`:
2071 for example, if :mdp:`init-lambda-state` is 5 and this parameter
2072 has a value of 2, energies for lambda points 3-7 will be calculated
2073 and writen out. A value of -1 means all lambda points will be
2074 written out. For normal BAR such as with :ref:`gmx bar`, a value of
2075 1 is sufficient, while for MBAR -1 should be used.
2080 the soft-core alpha parameter, a value of 0 results in linear
2081 interpolation of the LJ and Coulomb interactions
2086 the power of the radial term in the soft-core equation. Possible
2087 values are 6 and 48. 6 is more standard, and is the default. When
2088 48 is used, then sc-alpha should generally be much lower (between
2094 Whether to apply the soft core free energy interaction
2095 transformation to the Columbic interaction of a molecule. Default
2096 is no, as it is generally more efficient to turn off the Coulomic
2097 interactions linearly before turning off the van der Waals
2103 the power for lambda in the soft-core function, only the values 1
2109 the soft-core sigma for particles which have a C6 or C12 parameter
2110 equal to zero or a sigma smaller than :mdp:`sc-sigma`
2112 .. mdp:: couple-moltype
2114 Here one can supply a molecule type (as defined in the topology)
2115 for calculating solvation or coupling free energies. There is a
2116 special option ``system`` that couples all molecule types in the
2117 system. This can be useful for equilibrating a system starting from
2118 (nearly) random coordinates. :mdp:`free-energy` has to be turned
2119 on. The Van der Waals interactions and/or charges in this molecule
2120 type can be turned on or off between lambda=0 and lambda=1,
2121 depending on the settings of :mdp:`couple-lambda0` and
2122 :mdp:`couple-lambda1`. If you want to decouple one of several
2123 copies of a molecule, you need to copy and rename the molecule
2124 definition in the topology.
2126 .. mdp:: couple-lambda0
2128 .. mdp-value:: vdw-q
2130 all interactions are on at lambda=0
2134 the charges are zero (no Coulomb interactions) at lambda=0
2138 the Van der Waals interactions are turned at lambda=0; soft-core
2139 interactions will be required to avoid singularities
2143 the Van der Waals interactions are turned off and the charges
2144 are zero at lambda=0; soft-core interactions will be required to
2145 avoid singularities.
2147 .. mdp:: couple-lambda1
2149 analogous to :mdp:`couple-lambda1`, but for lambda=1
2151 .. mdp:: couple-intramol
2155 All intra-molecular non-bonded interactions for moleculetype
2156 :mdp:`couple-moltype` are replaced by exclusions and explicit
2157 pair interactions. In this manner the decoupled state of the
2158 molecule corresponds to the proper vacuum state without
2159 periodicity effects.
2163 The intra-molecular Van der Waals and Coulomb interactions are
2164 also turned on/off. This can be useful for partitioning
2165 free-energies of relatively large molecules, where the
2166 intra-molecular non-bonded interactions might lead to
2167 kinetically trapped vacuum conformations. The 1-4 pair
2168 interactions are not turned off.
2173 the frequency for writing dH/dlambda and possibly Delta H to
2174 dhdl.xvg, 0 means no ouput, should be a multiple of
2175 :mdp:`nstcalcenergy`.
2177 .. mdp:: dhdl-derivatives
2181 If yes (the default), the derivatives of the Hamiltonian with
2182 respect to lambda at each :mdp:`nstdhdl` step are written
2183 out. These values are needed for interpolation of linear energy
2184 differences with :ref:`gmx bar` (although the same can also be
2185 achieved with the right foreign lambda setting, that may not be as
2186 flexible), or with thermodynamic integration
2188 .. mdp:: dhdl-print-energy
2192 Include either the total or the potential energy in the dhdl
2193 file. Options are 'no', 'potential', or 'total'. This information
2194 is needed for later free energy analysis if the states of interest
2195 are at different temperatures. If all states are at the same
2196 temperature, this information is not needed. 'potential' is useful
2197 in case one is using ``mdrun -rerun`` to generate the ``dhdl.xvg``
2198 file. When rerunning from an existing trajectory, the kinetic
2199 energy will often not be correct, and thus one must compute the
2200 residual free energy from the potential alone, with the kinetic
2201 energy component computed analytically.
2203 .. mdp:: separate-dhdl-file
2207 The free energy values that are calculated (as specified with
2208 the foreign lambda and :mdp:`dhdl-derivatives` settings) are
2209 written out to a separate file, with the default name
2210 ``dhdl.xvg``. This file can be used directly with :ref:`gmx
2215 The free energy values are written out to the energy output file
2216 (``ener.edr``, in accumulated blocks at every :mdp:`nstenergy`
2217 steps), where they can be extracted with :ref:`gmx energy` or
2218 used directly with :ref:`gmx bar`.
2220 .. mdp:: dh-hist-size
2223 If nonzero, specifies the size of the histogram into which the
2224 Delta H values (specified with foreign lambda) and the derivative
2225 dH/dl values are binned, and written to ener.edr. This can be used
2226 to save disk space while calculating free energy differences. One
2227 histogram gets written for each foreign lambda and two for the
2228 dH/dl, at every :mdp:`nstenergy` step. Be aware that incorrect
2229 histogram settings (too small size or too wide bins) can introduce
2230 errors. Do not use histograms unless you're certain you need it.
2232 .. mdp:: dh-hist-spacing
2235 Specifies the bin width of the histograms, in energy units. Used in
2236 conjunction with :mdp:`dh-hist-size`. This size limits the
2237 accuracy with which free energies can be calculated. Do not use
2238 histograms unless you're certain you need it.
2241 Expanded Ensemble calculations
2242 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2244 .. mdp:: nstexpanded
2246 The number of integration steps beween attempted moves changing the
2247 system Hamiltonian in expanded ensemble simulations. Must be a
2248 multiple of :mdp:`nstcalcenergy`, but can be greater or less than
2255 No Monte Carlo in state space is performed.
2257 .. mdp-value:: metropolis-transition
2259 Uses the Metropolis weights to update the expanded ensemble
2260 weight of each state. Min{1,exp(-(beta_new u_new - beta_old
2263 .. mdp-value:: barker-transition
2265 Uses the Barker transition critera to update the expanded
2266 ensemble weight of each state i, defined by exp(-beta_new
2267 u_new)/(exp(-beta_new u_new)+exp(-beta_old u_old))
2269 .. mdp-value:: wang-landau
2271 Uses the Wang-Landau algorithm (in state space, not energy
2272 space) to update the expanded ensemble weights.
2274 .. mdp-value:: min-variance
2276 Uses the minimum variance updating method of Escobedo et al. to
2277 update the expanded ensemble weights. Weights will not be the
2278 free energies, but will rather emphasize states that need more
2279 sampling to give even uncertainty.
2281 .. mdp:: lmc-mc-move
2285 No Monte Carlo in state space is performed.
2287 .. mdp-value:: metropolis-transition
2289 Randomly chooses a new state up or down, then uses the
2290 Metropolis critera to decide whether to accept or reject:
2291 Min{1,exp(-(beta_new u_new - beta_old u_old)}
2293 .. mdp-value:: barker-transition
2295 Randomly chooses a new state up or down, then uses the Barker
2296 transition critera to decide whether to accept or reject:
2297 exp(-beta_new u_new)/(exp(-beta_new u_new)+exp(-beta_old u_old))
2299 .. mdp-value:: gibbs
2301 Uses the conditional weights of the state given the coordinate
2302 (exp(-beta_i u_i) / sum_k exp(beta_i u_i) to decide which state
2305 .. mdp-value:: metropolized-gibbs
2307 Uses the conditional weights of the state given the coordinate
2308 (exp(-beta_i u_i) / sum_k exp(beta_i u_i) to decide which state
2309 to move to, EXCLUDING the current state, then uses a rejection
2310 step to ensure detailed balance. Always more efficient that
2311 Gibbs, though only marginally so in many situations, such as
2312 when only the nearest neighbors have decent phase space
2318 random seed to use for Monte Carlo moves in state space. When
2319 :mdp:`lmc-seed` is set to -1, a pseudo random seed is us
2321 .. mdp:: mc-temperature
2323 Temperature used for acceptance/rejection for Monte Carlo moves. If
2324 not specified, the temperature of the simulation specified in the
2325 first group of :mdp:`ref-t` is used.
2330 The cutoff for the histogram of state occupancies to be reset, and
2331 the free energy incrementor to be changed from delta to delta *
2332 :mdp:`wl-scale`. If we define the Nratio = (number of samples at
2333 each histogram) / (average number of samples at each
2334 histogram). :mdp:`wl-ratio` of 0.8 means that means that the
2335 histogram is only considered flat if all Nratio > 0.8 AND
2336 simultaneously all 1/Nratio > 0.8.
2341 Each time the histogram is considered flat, then the current value
2342 of the Wang-Landau incrementor for the free energies is multiplied
2343 by :mdp:`wl-scale`. Value must be between 0 and 1.
2345 .. mdp:: init-wl-delta
2348 The initial value of the Wang-Landau incrementor in kT. Some value
2349 near 1 kT is usually most efficient, though sometimes a value of
2350 2-3 in units of kT works better if the free energy differences are
2353 .. mdp:: wl-oneovert
2356 Set Wang-Landau incrementor to scale with 1/(simulation time) in
2357 the large sample limit. There is significant evidence that the
2358 standard Wang-Landau algorithms in state space presented here
2359 result in free energies getting 'burned in' to incorrect values
2360 that depend on the initial state. when :mdp:`wl-oneovert` is true,
2361 then when the incrementor becomes less than 1/N, where N is the
2362 mumber of samples collected (and thus proportional to the data
2363 collection time, hence '1 over t'), then the Wang-Lambda
2364 incrementor is set to 1/N, decreasing every step. Once this occurs,
2365 :mdp:`wl-ratio` is ignored, but the weights will still stop
2366 updating when the equilibration criteria set in
2367 :mdp:`lmc-weights-equil` is achieved.
2369 .. mdp:: lmc-repeats
2372 Controls the number of times that each Monte Carlo swap type is
2373 performed each iteration. In the limit of large numbers of Monte
2374 Carlo repeats, then all methods converge to Gibbs sampling. The
2375 value will generally not need to be different from 1.
2377 .. mdp:: lmc-gibbsdelta
2380 Limit Gibbs sampling to selected numbers of neighboring states. For
2381 Gibbs sampling, it is sometimes inefficient to perform Gibbs
2382 sampling over all of the states that are defined. A positive value
2383 of :mdp:`lmc-gibbsdelta` means that only states plus or minus
2384 :mdp:`lmc-gibbsdelta` are considered in exchanges up and down. A
2385 value of -1 means that all states are considered. For less than 100
2386 states, it is probably not that expensive to include all states.
2388 .. mdp:: lmc-forced-nstart
2391 Force initial state space sampling to generate weights. In order to
2392 come up with reasonable initial weights, this setting allows the
2393 simulation to drive from the initial to the final lambda state,
2394 with :mdp:`lmc-forced-nstart` steps at each state before moving on
2395 to the next lambda state. If :mdp:`lmc-forced-nstart` is
2396 sufficiently long (thousands of steps, perhaps), then the weights
2397 will be close to correct. However, in most cases, it is probably
2398 better to simply run the standard weight equilibration algorithms.
2400 .. mdp:: nst-transition-matrix
2403 Frequency of outputting the expanded ensemble transition matrix. A
2404 negative number means it will only be printed at the end of the
2407 .. mdp:: symmetrized-transition-matrix
2410 Whether to symmetrize the empirical transition matrix. In the
2411 infinite limit the matrix will be symmetric, but will diverge with
2412 statistical noise for short timescales. Forced symmetrization, by
2413 using the matrix T_sym = 1/2 (T + transpose(T)), removes problems
2414 like the existence of (small magnitude) negative eigenvalues.
2416 .. mdp:: mininum-var-min
2419 The min-variance strategy (option of :mdp:`lmc-stats` is only
2420 valid for larger number of samples, and can get stuck if too few
2421 samples are used at each state. :mdp:`mininum-var-min` is the
2422 minimum number of samples that each state that are allowed before
2423 the min-variance strategy is activated if selected.
2425 .. mdp:: init-lambda-weights:
2427 The initial weights (free energies) used for the expanded ensemble
2428 states. Default is a vector of zero weights. format is similar to
2429 the lambda vector settings in :mdp:`fep-lambdas`, except the
2430 weights can be any floating point number. Units are kT. Its length
2431 must match the lambda vector lengths.
2433 .. mdp:: lmc-weights-equil
2437 Expanded ensemble weights continue to be updated throughout the
2442 The input expanded ensemble weights are treated as equilibrated,
2443 and are not updated throughout the simulation.
2445 .. mdp-value:: wl-delta
2447 Expanded ensemble weight updating is stopped when the
2448 Wang-Landau incrementor falls below this value.
2450 .. mdp-value:: number-all-lambda
2452 Expanded ensemble weight updating is stopped when the number of
2453 samples at all of the lambda states is greater than this value.
2455 .. mdp-value:: number-steps
2457 Expanded ensemble weight updating is stopped when the number of
2458 steps is greater than the level specified by this value.
2460 .. mdp-value:: number-samples
2462 Expanded ensemble weight updating is stopped when the number of
2463 total samples across all lambda states is greater than the level
2464 specified by this value.
2466 .. mdp-value:: count-ratio
2468 Expanded ensemble weight updating is stopped when the ratio of
2469 samples at the least sampled lambda state and most sampled
2470 lambda state greater than this value.
2472 .. mdp:: simulated-tempering
2475 Turn simulated tempering on or off. Simulated tempering is
2476 implemented as expanded ensemble sampling with different
2477 temperatures instead of different Hamiltonians.
2479 .. mdp:: sim-temp-low
2482 Low temperature for simulated tempering.
2484 .. mdp:: sim-temp-high
2487 High temperature for simulated tempering.
2489 .. mdp:: simulated-tempering-scaling
2491 Controls the way that the temperatures at intermediate lambdas are
2492 calculated from the :mdp:`temperature-lambdas` part of the lambda
2495 .. mdp-value:: linear
2497 Linearly interpolates the temperatures using the values of
2498 :mdp:`temperature-lambdas`, *i.e.* if :mdp:`sim-temp-low`
2499 =300, :mdp:`sim-temp-high` =400, then lambda=0.5 correspond to
2500 a temperature of 350. A nonlinear set of temperatures can always
2501 be implemented with uneven spacing in lambda.
2503 .. mdp-value:: geometric
2505 Interpolates temperatures geometrically between
2506 :mdp:`sim-temp-low` and :mdp:`sim-temp-high`. The i:th state
2507 has temperature :mdp:`sim-temp-low` * (:mdp:`sim-temp-high` /
2508 :mdp:`sim-temp-low`) raised to the power of
2509 (i/(ntemps-1)). This should give roughly equal exchange for
2510 constant heat capacity, though of course things simulations that
2511 involve protein folding have very high heat capacity peaks.
2513 .. mdp-value:: exponential
2515 Interpolates temperatures exponentially between
2516 :mdp:`sim-temp-low` and :mdp:`sim-temp-high`. The i:th state
2517 has temperature :mdp:`sim-temp-low` + (:mdp:`sim-temp-high` -
2518 :mdp:`sim-temp-low`)*((exp(:mdp:`temperature-lambdas`
2519 (i))-1)/(exp(1.0)-i)).
2527 groups for constant acceleration (*e.g.* ``Protein Sol``) all atoms
2528 in groups Protein and Sol will experience constant acceleration as
2529 specified in the :mdp:`accelerate` line
2534 acceleration for :mdp:`acc-grps`; x, y and z for each group
2535 (*e.g.* ``0.1 0.0 0.0 -0.1 0.0 0.0`` means that first group has
2536 constant acceleration of 0.1 nm ps-2 in X direction, second group
2541 Groups that are to be frozen (*i.e.* their X, Y, and/or Z position
2542 will not be updated; *e.g.* ``Lipid SOL``). :mdp:`freezedim`
2543 specifies for which dimension the freezing applies. To avoid
2544 spurious contibrutions to the virial and pressure due to large
2545 forces between completely frozen atoms you need to use energy group
2546 exclusions, this also saves computing time. Note that coordinates
2547 of frozen atoms are not scaled by pressure-coupling algorithms.
2551 dimensions for which groups in :mdp:`freezegrps` should be frozen,
2552 specify `Y` or `N` for X, Y and Z and for each group (*e.g.* ``Y Y
2553 N N N N`` means that particles in the first group can move only in
2554 Z direction. The particles in the second group can move in any
2557 .. mdp:: cos-acceleration
2560 the amplitude of the acceleration profile for calculating the
2561 viscosity. The acceleration is in the X-direction and the magnitude
2562 is :mdp:`cos-acceleration` cos(2 pi z/boxheight). Two terms are
2563 added to the energy file: the amplitude of the velocity profile and
2568 (0 0 0 0 0 0) \[nm ps-1\]
2569 The velocities of deformation for the box elements: a(x) b(y) c(z)
2570 b(x) c(x) c(y). Each step the box elements for which :mdp:`deform`
2571 is non-zero are calculated as: box(ts)+(t-ts)*deform, off-diagonal
2572 elements are corrected for periodicity. The coordinates are
2573 transformed accordingly. Frozen degrees of freedom are (purposely)
2574 also transformed. The time ts is set to t at the first step and at
2575 steps at which x and v are written to trajectory to ensure exact
2576 restarts. Deformation can be used together with semiisotropic or
2577 anisotropic pressure coupling when the appropriate
2578 compressibilities are set to zero. The diagonal elements can be
2579 used to strain a solid. The off-diagonal elements can be used to
2580 shear a solid or a liquid.
2586 .. mdp:: E-x ; E-y ; E-z
2588 If you want to use an electric field in a direction, enter 3
2589 numbers after the appropriate E-direction, the first number: the
2590 number of cosines, only 1 is implemented (with frequency 0) so
2591 enter 1, the second number: the strength of the electric field in V
2592 nm^-1, the third number: the phase of the cosine, you can enter any
2593 number here since a cosine of frequency zero has no phase.
2595 .. mdp:: E-xt; E-yt; E-zt:
2597 Here you can specify a pulsed alternating electric field. The field
2598 has the form of a gaussian laser pulse:
2600 E(t) = E0 exp ( -(t-t0)^2/(2 sigma^2) ) cos(omega (t-t0))
2602 For example, the four parameters for direction x are set in the
2603 three fields of :mdp:`E-x` and :mdp:`E-xt` like
2607 E-xt = omega t0 sigma
2609 In the special case that sigma = 0, the exponential term is omitted
2610 and only the cosine term is used.
2612 More details in Carl Caleman and David van der Spoel: Picosecond
2613 Melting of Ice by an Infrared Laser Pulse - A Simulation Study
2614 Angew. Chem. Intl. Ed. 47 pp. 14 17-1420 (2008)
2618 Mixed quantum/classical molecular dynamics
2619 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2629 Do a QM/MM simulation. Several groups can be described at
2630 different QM levels separately. These are specified in the
2631 :mdp:`QMMM-grps` field separated by spaces. The level of *ab
2632 initio* theory at which the groups are described is specified by
2633 :mdp:`QMmethod` and :mdp:`QMbasis` Fields. Describing the
2634 groups at different levels of theory is only possible with the
2635 ONIOM QM/MM scheme, specified by :mdp:`QMMMscheme`.
2639 groups to be descibed at the QM level
2643 .. mdp-value:: normal
2645 normal QM/MM. There can only be one :mdp:`QMMM-grps` that is
2646 modelled at the :mdp:`QMmethod` and :mdp:`QMbasis` level of
2647 *ab initio* theory. The rest of the system is described at the
2648 MM level. The QM and MM subsystems interact as follows: MM point
2649 charges are included in the QM one-electron hamiltonian and all
2650 Lennard-Jones interactions are described at the MM level.
2652 .. mdp-value:: ONIOM
2654 The interaction between the subsystem is described using the
2655 ONIOM method by Morokuma and co-workers. There can be more than
2656 one :mdp:`QMMM-grps` each modeled at a different level of QM
2657 theory (:mdp:`QMmethod` and :mdp:`QMbasis`).
2662 Method used to compute the energy and gradients on the QM
2663 atoms. Available methods are AM1, PM3, RHF, UHF, DFT, B3LYP, MP2,
2664 CASSCF, and MMVB. For CASSCF, the number of electrons and orbitals
2665 included in the active space is specified by :mdp:`CASelectrons`
2666 and :mdp:`CASorbitals`.
2671 Basis set used to expand the electronic wavefuntion. Only Gaussian
2672 basis sets are currently available, *i.e.* ``STO-3G, 3-21G, 3-21G*,
2673 3-21+G*, 6-21G, 6-31G, 6-31G*, 6-31+G*,`` and ``6-311G``.
2678 The total charge in `e` of the :mdp:`QMMM-grps`. In case there are
2679 more than one :mdp:`QMMM-grps`, the total charge of each ONIOM
2680 layer needs to be specified separately.
2685 The multiplicity of the :mdp:`QMMM-grps`. In case there are more
2686 than one :mdp:`QMMM-grps`, the multiplicity of each ONIOM layer
2687 needs to be specified separately.
2689 .. mdp:: CASorbitals
2692 The number of orbitals to be included in the active space when
2693 doing a CASSCF computation.
2695 .. mdp:: CASelectrons
2698 The number of electrons to be included in the active space when
2699 doing a CASSCF computation.
2705 No surface hopping. The system is always in the electronic
2710 Do a QM/MM MD simulation on the excited state-potential energy
2711 surface and enforce a *diabatic* hop to the ground-state when
2712 the system hits the conical intersection hyperline in the course
2713 the simulation. This option only works in combination with the
2720 .. mdp:: implicit-solvent
2728 Do a simulation with implicit solvent using the Generalized Born
2729 formalism. Three different methods for calculating the Born
2730 radii are available, Still, HCT and OBC. These are specified
2731 with the :mdp:`gb-algorithm` field. The non-polar solvation is
2732 specified with the :mdp:`sa-algorithm` field.
2734 .. mdp:: gb-algorithm
2736 .. mdp-value:: Still
2738 Use the Still method to calculate the Born radii
2742 Use the Hawkins-Cramer-Truhlar method to calculate the Born
2747 Use the Onufriev-Bashford-Case method to calculate the Born
2753 Frequency to (re)-calculate the Born radii. For most practial
2754 purposes, setting a value larger than 1 violates energy
2755 conservation and leads to unstable trajectories.
2760 Cut-off for the calculation of the Born radii. Currently must be
2763 .. mdp:: gb-epsilon-solvent
2766 Dielectric constant for the implicit solvent
2768 .. mdp:: gb-saltconc
2771 Salt concentration for implicit solvent models, currently not used
2773 .. mdp:: gb-obc-alpha
2774 .. mdp:: gb-obc-beta
2775 .. mdp:: gb-obc-gamma
2777 Scale factors for the OBC model. Default values of 1, 0.78 and 4.85
2778 respectively are for OBC(II). Values for OBC(I) are 0.8, 0 and 2.91
2781 .. mdp:: gb-dielectric-offset
2784 Distance for the di-electric offset when calculating the Born
2785 radii. This is the offset between the center of each atom the
2786 center of the polarization energy for the corresponding atom
2788 .. mdp:: sa-algorithm
2790 .. mdp-value:: Ace-approximation
2792 Use an Ace-type approximation
2796 No non-polar solvation calculation done. For GBSA only the polar
2797 part gets calculated
2799 .. mdp:: sa-surface-tension
2802 Default value for surface tension with SA algorithms. The default
2803 value is -1; Note that if this default value is not changed it will
2804 be overridden by :ref:`gmx grompp` using values that are specific
2805 for the choice of radii algorithm (0.0049 kcal/mol/Angstrom^2 for
2806 Still, 0.0054 kcal/mol/Angstrom2 for HCT/OBC) Setting it to 0 will
2807 while using an sa-algorithm other than None means no non-polar
2808 calculations are done.
2811 Adaptive Resolution Simulation
2812 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2817 Decide whether the AdResS feature is turned on.
2819 .. mdp:: adress-type
2823 Do an AdResS simulation with weight equal 1, which is equivalent
2824 to an explicit (normal) MD simulation. The difference to
2825 disabled AdResS is that the AdResS variables are still read-in
2826 and hence are defined.
2828 .. mdp-value:: Constant
2830 Do an AdResS simulation with a constant weight,
2831 :mdp:`adress-const-wf` defines the value of the weight
2833 .. mdp-value:: XSplit
2835 Do an AdResS simulation with simulation box split in
2836 x-direction, so basically the weight is only a function of the x
2837 coordinate and all distances are measured using the x coordinate
2840 .. mdp-value:: Sphere
2842 Do an AdResS simulation with spherical explicit zone.
2844 .. mdp:: adress-const-wf
2847 Provides the weight for a constant weight simulation
2848 (:mdp:`adress-type` =Constant)
2850 .. mdp:: adress-ex-width
2853 Width of the explicit zone, measured from
2854 :mdp:`adress-reference-coords`.
2856 .. mdp:: adress-hy-width
2859 Width of the hybrid zone.
2861 .. mdp:: adress-reference-coords
2864 Position of the center of the explicit zone. Periodic boundary
2865 conditions apply for measuring the distance from it.
2867 .. mdp:: adress-cg-grp-names
2869 The names of the coarse-grained energy groups. All other energy
2870 groups are considered explicit and their interactions will be
2871 automatically excluded with the coarse-grained groups.
2873 .. mdp:: adress-site
2875 The mapping point from which the weight is calculated.
2879 The weight is calculated from the center of mass of each charge group.
2883 The weight is calculated from the center of geometry of each charge group.
2887 The weight is calculated from the position of 1st atom of each charge group.
2889 .. mdp-value:: AtomPerAtom
2891 The weight is calculated from the position of each individual atom.
2893 .. mdp:: adress-interface-correction
2897 Do not apply any interface correction.
2899 .. mdp-value:: thermoforce
2901 Apply thermodynamic force interface correction. The table can be
2902 specified using the ``-tabletf`` option of :ref:`gmx mdrun`. The
2903 table should contain the potential and force (acting on
2904 molecules) as function of the distance from
2905 :mdp:`adress-reference-coords`.
2907 .. mdp:: adress-tf-grp-names
2909 The names of the energy groups to which the thermoforce is applied
2910 if enabled in :mdp:`adress-interface-correction`. If no group is
2911 given the default table is applied.
2913 .. mdp:: adress-ex-forcecap
2916 Cap the force in the hybrid region, useful for big molecules. 0
2917 disables force capping.
2920 User defined thingies
2921 ^^^^^^^^^^^^^^^^^^^^^
2925 .. mdp:: userint1 (0)
2926 .. mdp:: userint2 (0)
2927 .. mdp:: userint3 (0)
2928 .. mdp:: userint4 (0)
2929 .. mdp:: userreal1 (0)
2930 .. mdp:: userreal2 (0)
2931 .. mdp:: userreal3 (0)
2932 .. mdp:: userreal4 (0)
2934 These you can use if you modify code. You can pass integers and
2935 reals and groups to your subroutine. Check the inputrec definition
2936 in ``src/gromacs/legacyheaders/types/inputrec.h``
2938 .. _reference manual: gmx-manual-parent-dir_