1 "LAMMPS WWW Site"_lws - "LAMMPS Documentation"_ld - "LAMMPS Commands"_lc :c
3 :link(lws,http://lammps.sandia.gov)
5 :link(lc,Section_commands.html#comm)
13 fix ID id-group gle Ns Tstart Tstop seed Amatrix \[noneq Cmatrix\] \[every stride\] :pre
15 ID, group-ID are documented in "fix"_fix.html command :ulb,l
16 gle = style name of this fix command :l
17 Ns = number of additional fictitious momenta :l
18 Tstart, Tstop = temperature ramp during the run :l
19 Amatrix = file to read the drift matrix A from :l
20 seed = random number seed to use for generating noise (positive integer) :l
21 zero or more keyword/value pairs may be appended :l
22 keyword = {noneq} or {every}
23 {noneq} Cmatrix = file to read the non-equilibrium covariance matrix from
24 {every} stride = apply the GLE once every time steps. Reduces the accuracy
25 of the integration of the GLE, but has *no effect* on the accuracy of equilibrium
26 sampling. It might change sampling properties when used together with {noneq}. :pre
31 fix 3 boundary gle 6 300 300 31415 smart.A
32 fix 1 all gle 6 300 300 31415 qt-300k.A noneq qt-300k.C :pre
36 Apply a Generalized Langevin Equation (GLE) thermostat as described
37 in "(Ceriotti)"_#Ceriotti. The formalism allows one to obtain a number
38 of different effects ranging from efficient sampling of all
39 vibrational modes in the system to inexpensive (approximate)
40 modelling of nuclear quantum effects. Contrary to
41 "fix langevin"_fix_langevin.html, this fix performs both
42 thermostatting and evolution of the Hamiltonian equations of motion, so it
43 should not be used together with "fix nve"_fix_nve.html -- at least not
44 on the same atom groups.
46 Each degree of freedom in the thermostatted group is supplemented
47 with Ns additional degrees of freedom s, and the equations of motion
51 d(p,s)/dt=(F,0) - A(p,s) + B dW/dt :pre
53 where F is the physical force, A is the drift matrix (that generalizes
54 the friction in Langevin dynamics), B is the diffusion term and dW/dt
55 un-correlated Gaussian random forces. The A matrix couples the physical
56 (q,p) dynamics with that of the additional degrees of freedom,
57 and makes it possible to obtain effectively a history-dependent
58 noise and friction kernel.
60 The drift matrix should be given as an external file {Afile},
61 as a (Ns+1 x Ns+1) matrix in inverse time units. Matrices that are
62 optimal for a given application and the system of choice can be
63 obtained from "(GLE4MD)"_#GLE4MD.
65 Equilibrium sampling a temperature T is obtained by specifiying the
66 target value as the {Tstart} and {Tstop} arguments, so that the diffusion
67 matrix that gives canonical sampling for a given A is computed automatically.
68 However, the GLE framework also allow for non-equilibrium sampling, that
69 can be used for instance to model inexpensively zero-point energy
70 effects "(Ceriotti2)"_#Ceriotti2. This is achieved specifying the
71 {noneq} keyword followed by the name of the file that contains the
72 static covariance matrix for the non-equilibrium dynamics.
74 Since integrating GLE dynamics can be costly when used together with
75 simple potentials, one can use the {every} optional keyword to
76 apply the Langevin terms only once every several MD steps, in a
77 multiple time-step fashion. This should be used with care when doing
78 non-equilibrium sampling, but should have no effect on equilibrium
79 averages when using canonical sampling.
81 The random number {seed} must be a positive integer. A Marsaglia random
82 number generator is used. Each processor uses the input seed to
83 generate its own unique seed and its own stream of random numbers.
84 Thus the dynamics of the system will not be identical on two runs on
85 different numbers of processors.
87 Note also that the Generalized Langevin Dynamics scheme that is
88 implemented by the "fix gld"_fix_gld.html scheme is closely related
89 to the present one. In fact, it should be always possible to cast the
90 Prony series form of the memory kernel used by GLD into an appropriate
91 input matrix for "fix gle"_fix_gle.html. While the GLE scheme is more
92 general, the form used by "fix gld"_fix_gld.html can be more directly
93 related to the representation of an implicit solvent environment.
95 [Restart, fix_modify, output, run start/stop, minimize info:]
97 The instantaneous values of the extended variables are written to
98 "binary restart files"_restart.html. Because the state of the random
99 number generator is not saved in restart files, this means you cannot
100 do "exact" restarts with this fix, where the simulation continues on
101 the same as if no restart had taken place. However, in a statistical
102 sense, a restarted simulation should produce the same behavior.
103 Note however that you should use a different seed each time you
104 restart, otherwise the same sequence of random numbers will be used
105 each time, which might lead to stochastic synchronization and
106 subtle artefacts in the sampling.
108 This fix can ramp its target temperature over multiple runs, using the
109 {start} and {stop} keywords of the "run"_run.html command. See the
110 "run"_run.html command for details of how to do this.
112 The "fix_modify"_fix_modify.html {energy} option is supported by this
113 fix to add the energy change induced by Langevin thermostatting to the
114 system's potential energy as part of "thermodynamic
115 output"_thermo_style.html.
117 This fix computes a global scalar which can be accessed by various
118 "output commands"_Section_howto.html#howto_15. The scalar is the
119 cummulative energy change due to this fix. The scalar value
120 calculated by this fix is "extensive".
124 The GLE thermostat in its current implementation should not be used
125 with rigid bodies, SHAKE or RATTLE. It is expected that all the
126 thermostatted degrees of freedom are fully flexible, and the sampled
127 ensemble will not be correct otherwise.
129 In order to perform constant-pressure simulations please use
130 "fix press/berendsen"_fix_press_berendsen.html, rather than
131 "fix npt"_fix_nh.html, to avoid duplicate integration of the
134 This fix is part of the USER-MISC package. It is only enabled if LAMMPS
135 was built with that package. See the "Making
136 LAMMPS"_Section_start.html#start_3 section for more info.
140 "fix nvt"_fix_nh.html, "fix temp/rescale"_fix_temp_rescale.html, "fix
141 viscous"_fix_viscous.html, "fix nvt"_fix_nh.html, "pair_style
142 dpd/tstat"_pair_dpd.html, "fix gld"_fix_gld.html
147 [(Ceriotti)] Ceriotti, Bussi and Parrinello, J Chem Theory Comput 6,
151 [(GLE4MD)] "http://epfl-cosmo.github.io/gle4md/"_http://epfl-cosmo.github.io/gle4md/
154 [(Ceriotti2)] Ceriotti, Bussi and Parrinello, Phys Rev Lett 103,