Line data Source code
1 : /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 : Copyright (c) 2011-2023 The plumed team
3 : (see the PEOPLE file at the root of the distribution for a list of names)
4 :
5 : See http://www.plumed.org for more information.
6 :
7 : This file is part of plumed, version 2.
8 :
9 : plumed is free software: you can redistribute it and/or modify
10 : it under the terms of the GNU Lesser General Public License as published by
11 : the Free Software Foundation, either version 3 of the License, or
12 : (at your option) any later version.
13 :
14 : plumed is distributed in the hope that it will be useful,
15 : but WITHOUT ANY WARRANTY; without even the implied warranty of
16 : MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 : GNU Lesser General Public License for more details.
18 :
19 : You should have received a copy of the GNU Lesser General Public License
20 : along with plumed. If not, see <http://www.gnu.org/licenses/>.
21 : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
22 : #include "Bias.h"
23 : #include "core/ActionRegister.h"
24 : #include "core/ActionSet.h"
25 : #include "core/PlumedMain.h"
26 : #include "core/FlexibleBin.h"
27 : #include "tools/Exception.h"
28 : #include "tools/Grid.h"
29 : #include "tools/Matrix.h"
30 : #include "tools/OpenMP.h"
31 : #include "tools/Random.h"
32 : #include "tools/File.h"
33 : #include "tools/Communicator.h"
34 : #include <ctime>
35 : #include <numeric>
36 :
37 : namespace PLMD {
38 : namespace bias {
39 :
40 : //+PLUMEDOC BIAS METAD
41 : /*
42 : Used to performed metadynamics on one or more collective variables.
43 :
44 : In a metadynamics simulations a history dependent bias composed of
45 : intermittently added Gaussian functions is added to the potential \cite metad.
46 :
47 : \f[
48 : V(\vec{s},t) = \sum_{ k \tau < t} W(k \tau)
49 : \exp\left(
50 : -\sum_{i=1}^{d} \frac{(s_i-s_i^{(0)}(k \tau))^2}{2\sigma_i^2}
51 : \right).
52 : \f]
53 :
54 : This potential forces the system away from the kinetic traps in the potential energy surface
55 : and out into the unexplored parts of the energy landscape. Information on the Gaussian
56 : functions from which this potential is composed is output to a file called HILLS, which
57 : is used both the restart the calculation and to reconstruct the free energy as a function of the CVs.
58 : The free energy can be reconstructed from a metadynamics calculation because the final bias is given
59 : by:
60 :
61 : \f[
62 : V(\vec{s}) = -F(\vec{s})
63 : \f]
64 :
65 : During post processing the free energy can be calculated in this way using the \ref sum_hills
66 : utility.
67 :
68 : In the simplest possible implementation of a metadynamics calculation the expense of a metadynamics
69 : calculation increases with the length of the simulation as one has to, at every step, evaluate
70 : the values of a larger and larger number of Gaussian kernels. To avoid this issue you can
71 : store the bias on a grid. This approach is similar to that proposed in \cite babi08jcp but has the
72 : advantage that the grid spacing is independent on the Gaussian width.
73 : Notice that you should provide the grid boundaries (GRID_MIN and GRID_MAX) and either the number of bins
74 : for every collective variable (GRID_BIN) or the desired grid spacing (GRID_SPACING).
75 : In case you provide both PLUMED will use the most conservative choice (highest number of bins) for each dimension.
76 : In case you do not provide any information about bin size (neither GRID_BIN nor GRID_SPACING)
77 : PLUMED will use 1/5 of the Gaussian width (SIGMA) as grid spacing if the width is fixed or 1/5 of the minimum
78 : Gaussian width (SIGMA_MIN) if the width is variable. This default choice should be reasonable for most applications.
79 :
80 : Alternatively to the use of grids, it is possible to use a neighbor list to decrease the cost of evaluating the bias,
81 : this can be enabled using NLIST. NLIST can be beneficial with more than 2 collective variables, where GRID becomes
82 : expensive and memory consuming. The neighbor list will be updated everytime the CVs go farther than a cut-off value
83 : from the position they were at last neighbor list update. Gaussians are added to the neigbhor list if their center
84 : is within 6.*DP2CUTOFF*sigma*sigma. While the list is updated if the CVs are farther from the center than 0.5 of the
85 : standard deviation of the Gaussian center distribution of the list. These parameters (6 and 0.5) can be modified using
86 : NLIST_PARAMETERS. Note that the use of neighbor list does not provide the exact bias.
87 :
88 : Metadynamics can be restarted either from a HILLS file as well as from a GRID, in this second
89 : case one can first save a GRID using GRID_WFILE (and GRID_WSTRIDE) and at a later stage read
90 : it using GRID_RFILE.
91 :
92 : The work performed by the METAD bias can be calculated using CALC_WORK, note that this is expensive when not using grids.
93 :
94 : Another option that is available in plumed is well-tempered metadynamics \cite Barducci:2008. In this
95 : variant of metadynamics the heights of the Gaussian hills are scaled at each step so the bias is now
96 : given by:
97 :
98 : \f[
99 : V({s},t)= \sum_{t'=0,\tau_G,2\tau_G,\dots}^{t'<t} W e^{-V({s}({q}(t'),t')/\Delta T} \exp\left(
100 : -\sum_{i=1}^{d} \frac{(s_i({q})-s_i({q}(t'))^2}{2\sigma_i^2}
101 : \right),
102 : \f]
103 :
104 : This method ensures that the bias converges more smoothly. It should be noted that, in the case of well-tempered metadynamics, in
105 : the output printed the Gaussian height is re-scaled using the bias factor.
106 : Also notice that with well-tempered metadynamics the HILLS file does not contain the bias,
107 : but the negative of the free-energy estimate. This choice has the advantage that
108 : one can restart a simulation using a different value for the \f$\Delta T\f$. The applied bias will be scaled accordingly.
109 :
110 : Note that you can use here also the flexible Gaussian approach \cite Branduardi:2012dl
111 : in which you can adapt the Gaussian to the extent of Cartesian space covered by a variable or
112 : to the space in collective variable covered in a given time. In this case the width of the deposited
113 : Gaussian potential is denoted by one value only that is a Cartesian space (ADAPTIVE=GEOM) or a time
114 : (ADAPTIVE=DIFF). Note that a specific integration technique for the deposited Gaussian kernels
115 : should be used in this case. Check the documentation for utility sum_hills.
116 :
117 : With the keyword INTERVAL one changes the metadynamics algorithm setting the bias force equal to zero
118 : outside boundary \cite baftizadeh2012protein. If, for example, metadynamics is performed on a CV s and one is interested only
119 : to the free energy for s > boundary, the history dependent potential is still updated according to the above
120 : equations but the metadynamics force is set to zero for s < boundary. Notice that Gaussian kernels are added also
121 : if s < boundary, as the tails of these Gaussian kernels influence VG in the relevant region s > boundary. In this way, the
122 : force on the system in the region s > boundary comes from both metadynamics and the force field, in the region
123 : s < boundary only from the latter. This approach allows obtaining a history-dependent bias potential VG that
124 : fluctuates around a stable estimator, equal to the negative of the free energy far enough from the
125 : boundaries. Note that:
126 : - It works only for one-dimensional biases;
127 : - It works both with and without GRID;
128 : - The interval limit boundary in a region where the free energy derivative is not large;
129 : - If in the region outside the limit boundary the system has a free energy minimum, the INTERVAL keyword should
130 : be used together with a \ref UPPER_WALLS or \ref LOWER_WALLS at boundary.
131 :
132 : As a final note, since version 2.0.2 when the system is outside of the selected interval the force
133 : is set to zero and the bias value to the value at the corresponding boundary. This allows acceptances
134 : for replica exchange methods to be computed correctly.
135 :
136 : Multiple walkers \cite multiplewalkers can also be used. See below the examples.
137 :
138 :
139 : The \f$c(t)\f$ reweighting factor can also be calculated on the fly using the equations
140 : presented in \cite Tiwary_jp504920s.
141 : The expression used to calculate \f$c(t)\f$ follows directly from Eq. 3 in \cite Tiwary_jp504920s,
142 : where \f$F(\vec{s})=-\gamma/(\gamma-1) V(\vec{s})\f$.
143 : This gives smoother results than equivalent Eqs. 13 and Eqs. 14 in that paper.
144 : The \f$c(t)\f$ is given by the rct component while the bias
145 : normalized by \f$c(t)\f$ is given by the rbias component (rbias=bias-rct) which can be used
146 : to obtain a reweighted histogram.
147 : The calculation of \f$c(t)\f$ is enabled by using the keyword CALC_RCT.
148 : By default \f$c(t)\f$ is updated every time the bias changes, but if this slows down the simulation
149 : the keyword RCT_USTRIDE can be set to a value higher than 1.
150 : This option requires that a grid is used.
151 :
152 : Additional material and examples can be also found in the tutorials:
153 :
154 : - \ref lugano-3
155 :
156 : Concurrent metadynamics
157 : as done e.g. in Ref. \cite gil2015enhanced . This indeed can be obtained by using the METAD
158 : action multiple times in the same input file.
159 :
160 : \par Examples
161 :
162 : The following input is for a standard metadynamics calculation using as
163 : collective variables the distance between atoms 3 and 5
164 : and the distance between atoms 2 and 4. The value of the CVs and
165 : the metadynamics bias potential are written to the COLVAR file every 100 steps.
166 : \plumedfile
167 : DISTANCE ATOMS=3,5 LABEL=d1
168 : DISTANCE ATOMS=2,4 LABEL=d2
169 : METAD ARG=d1,d2 SIGMA=0.2,0.2 HEIGHT=0.3 PACE=500 LABEL=restraint
170 : PRINT ARG=d1,d2,restraint.bias STRIDE=100 FILE=COLVAR
171 : \endplumedfile
172 : (See also \ref DISTANCE \ref PRINT).
173 :
174 : \par
175 : If you use adaptive Gaussian kernels, with diffusion scheme where you use
176 : a Gaussian that should cover the space of 20 time steps in collective variables.
177 : Note that in this case the histogram correction is needed when summing up hills.
178 : \plumedfile
179 : DISTANCE ATOMS=3,5 LABEL=d1
180 : DISTANCE ATOMS=2,4 LABEL=d2
181 : METAD ARG=d1,d2 SIGMA=20 HEIGHT=0.3 PACE=500 LABEL=restraint ADAPTIVE=DIFF
182 : PRINT ARG=d1,d2,restraint.bias STRIDE=100 FILE=COLVAR
183 : \endplumedfile
184 :
185 : \par
186 : If you use adaptive Gaussian kernels, with geometrical scheme where you use
187 : a Gaussian that should cover the space of 0.05 nm in Cartesian space.
188 : Note that in this case the histogram correction is needed when summing up hills.
189 : \plumedfile
190 : DISTANCE ATOMS=3,5 LABEL=d1
191 : DISTANCE ATOMS=2,4 LABEL=d2
192 : METAD ARG=d1,d2 SIGMA=0.05 HEIGHT=0.3 PACE=500 LABEL=restraint ADAPTIVE=GEOM
193 : PRINT ARG=d1,d2,restraint.bias STRIDE=100 FILE=COLVAR
194 : \endplumedfile
195 :
196 : \par
197 : When using adaptive Gaussian kernels you might want to limit how the hills width can change.
198 : You can use SIGMA_MIN and SIGMA_MAX keywords.
199 : The sigmas should specified in terms of CV so you should use the CV units.
200 : Note that if you use a negative number, this means that the limit is not set.
201 : Note also that in this case the histogram correction is needed when summing up hills.
202 : \plumedfile
203 : DISTANCE ATOMS=3,5 LABEL=d1
204 : DISTANCE ATOMS=2,4 LABEL=d2
205 : METAD ...
206 : ARG=d1,d2 SIGMA=0.05 HEIGHT=0.3 PACE=500 LABEL=restraint ADAPTIVE=GEOM
207 : SIGMA_MIN=0.2,0.1 SIGMA_MAX=0.5,1.0
208 : ... METAD
209 : PRINT ARG=d1,d2,restraint.bias STRIDE=100 FILE=COLVAR
210 : \endplumedfile
211 :
212 : \par
213 : Multiple walkers can be also use as in \cite multiplewalkers
214 : These are enabled by setting the number of walker used, the id of the
215 : current walker which interprets the input file, the directory where the
216 : hills containing files resides, and the frequency to read the other walkers.
217 : Here is an example
218 : \plumedfile
219 : DISTANCE ATOMS=3,5 LABEL=d1
220 : METAD ...
221 : ARG=d1 SIGMA=0.05 HEIGHT=0.3 PACE=500 LABEL=restraint
222 : WALKERS_N=10
223 : WALKERS_ID=3
224 : WALKERS_DIR=../
225 : WALKERS_RSTRIDE=100
226 : ... METAD
227 : \endplumedfile
228 : where WALKERS_N is the total number of walkers, WALKERS_ID is the
229 : id of the present walker (starting from 0 ) and the WALKERS_DIR is the directory
230 : where all the walkers are located. WALKERS_RSTRIDE is the number of step between
231 : one update and the other. Since version 2.2.5, hills files are automatically
232 : flushed every WALKERS_RSTRIDE steps.
233 :
234 : \par
235 : The \f$c(t)\f$ reweighting factor can be calculated on the fly using the equations
236 : presented in \cite Tiwary_jp504920s as described above.
237 : This is enabled by using the keyword CALC_RCT,
238 : and can be done only if the bias is defined on a grid.
239 : \plumedfile
240 : phi: TORSION ATOMS=1,2,3,4
241 : psi: TORSION ATOMS=5,6,7,8
242 :
243 : METAD ...
244 : LABEL=metad
245 : ARG=phi,psi SIGMA=0.20,0.20 HEIGHT=1.20 BIASFACTOR=5 TEMP=300.0 PACE=500
246 : GRID_MIN=-pi,-pi GRID_MAX=pi,pi GRID_BIN=150,150
247 : CALC_RCT
248 : RCT_USTRIDE=10
249 : ... METAD
250 : \endplumedfile
251 : Here we have asked that the calculation is performed every 10 hills deposition by using
252 : RCT_USTRIDE keyword. If this keyword is not given, the calculation will
253 : by default be performed every time the bias changes. The \f$c(t)\f$ reweighting factor will be given
254 : in the rct component while the instantaneous value of the bias potential
255 : normalized using the \f$c(t)\f$ reweighting factor is given in the rbias component
256 : [rbias=bias-rct] which can be used to obtain a reweighted histogram or
257 : free energy surface using the \ref HISTOGRAM analysis.
258 :
259 : \par
260 : The kinetics of the transitions between basins can also be analyzed on the fly as
261 : in \cite PRL230602. The flag ACCELERATION turn on accumulation of the acceleration
262 : factor that can then be used to determine the rate. This method can be used together
263 : with \ref COMMITTOR analysis to stop the simulation when the system get to the target basin.
264 : It must be used together with Well-Tempered Metadynamics. If restarting from a previous
265 : metadynamics you need to use the ACCELERATION_RFILE keyword to give the name of the
266 : data file from which the previous value of the acceleration factor should be read, otherwise the
267 : calculation of the acceleration factor will be wrong.
268 :
269 : \par
270 : By using the flag FREQUENCY_ADAPTIVE the frequency adaptive scheme introduced in \cite Wang-JCP-2018
271 : is turned on. The frequency for hill addition then changes dynamically based on the acceleration factor
272 : according to the following equation
273 : \f[
274 : \tau_{\mathrm{dep}}(t) =
275 : \min\left[
276 : \tau_0 \cdot
277 : \max\left[\frac{\alpha(t)}{\theta},1\right]
278 : ,\tau_{c}
279 : \right]
280 : \f]
281 : where \f$\tau_0\f$ is the initial hill addition frequency given by the PACE keyword,
282 : \f$\tau_{c}\f$ is the maximum allowed frequency given by the FA_MAX_PACE keyword,
283 : \f$\alpha(t)\f$ is the instantaneous acceleration factor at time \f$t\f$,
284 : and \f$\theta\f$ is a threshold value that acceleration factor has to reach before
285 : triggering a change in the hill addition frequency given by the FA_MIN_ACCELERATION keyword.
286 : The frequency for updating the hill addition frequency according to this equation is
287 : given by the FA_UPDATE_FREQUENCY keyword, by default it is the same as the value given
288 : in PACE. The hill hill addition frequency increase monotonously such that if the
289 : instantaneous acceleration factor is lower than in the previous updating step the
290 : previous \f$\tau_{\mathrm{dep}}\f$ is kept rather than updating it to a lower value.
291 : The instantaneous hill addition frequency \f$\tau_{\mathrm{dep}}(t)\f$ is outputted
292 : to pace component. Note that if restarting from a previous metadynamics run you need to
293 : use the ACCELERATION_RFILE keyword to read in the acceleration factors from the
294 : previous run, otherwise the hill addition frequency will start from the initial
295 : frequency.
296 :
297 :
298 : \par
299 : You can also provide a target distribution using the keyword TARGET
300 : \cite white2015designing
301 : \cite marinelli2015ensemble
302 : \cite gil2016empirical
303 : The TARGET should be a grid containing a free-energy (i.e. the -\f$k_B\f$T*log of the desired target distribution).
304 : Gaussian kernels will then be scaled by a factor
305 : \f[
306 : e^{\beta(\tilde{F}(s)-\tilde{F}_{max})}
307 : \f]
308 : Here \f$\tilde{F}(s)\f$ is the free energy defined on the grid and \f$\tilde{F}_{max}\f$ its maximum value.
309 : Notice that we here used the maximum value as in ref \cite gil2016empirical
310 : This choice allows to avoid exceedingly large Gaussian kernels to be added. However,
311 : it could make the Gaussian too small. You should always choose carefully the HEIGHT parameter
312 : in this case.
313 : The grid file should be similar to other PLUMED grid files in that it should contain
314 : both the target free-energy and its derivatives.
315 :
316 : Notice that if you wish your simulation to converge to the target free energy you should use
317 : the DAMPFACTOR command to provide a global tempering \cite dama2014well
318 : Alternatively, if you use a BIASFACTOR your simulation will converge to a free
319 : energy that is a linear combination of the target free energy and of the intrinsic free energy
320 : determined by the original force field.
321 :
322 : \plumedfile
323 : DISTANCE ATOMS=3,5 LABEL=d1
324 : METAD ...
325 : LABEL=t1
326 : ARG=d1 SIGMA=0.05 TAU=200 DAMPFACTOR=100 PACE=250
327 : GRID_MIN=1.14 GRID_MAX=1.32 GRID_BIN=6
328 : TARGET=dist.grid
329 : ... METAD
330 :
331 : PRINT ARG=d1,t1.bias STRIDE=100 FILE=COLVAR
332 : \endplumedfile
333 :
334 : The file dist.dat for this calculation would read:
335 :
336 : \auxfile{dist.grid}
337 : #! FIELDS d1 t1.target der_d1
338 : #! SET min_d1 1.14
339 : #! SET max_d1 1.32
340 : #! SET nbins_d1 6
341 : #! SET periodic_d1 false
342 : 1.1400 0.0031 0.1101
343 : 1.1700 0.0086 0.2842
344 : 1.2000 0.0222 0.6648
345 : 1.2300 0.0521 1.4068
346 : 1.2600 0.1120 2.6873
347 : 1.2900 0.2199 4.6183
348 : 1.3200 0.3948 7.1055
349 : \endauxfile
350 :
351 : Notice that BIASFACTOR can also be chosen as equal to 1. In this case one will perform
352 : unbiased sampling. Instead of using HEIGHT, one should provide the TAU parameter.
353 : \plumedfile
354 : d: DISTANCE ATOMS=3,5
355 : METAD ARG=d SIGMA=0.1 TAU=4.0 TEMP=300 PACE=100 BIASFACTOR=1.0
356 : \endplumedfile
357 : The HILLS file obtained will still work with `plumed sum_hills` so as to plot a free-energy.
358 : The case where this makes sense is probably that of RECT simulations.
359 :
360 : Regarding RECT simulations, you can also use the RECT keyword so as to avoid using multiple input files.
361 : For instance, a single input file will be
362 : \plumedfile
363 : d: DISTANCE ATOMS=3,5
364 : METAD ARG=d SIGMA=0.1 TAU=4.0 TEMP=300 PACE=100 RECT=1.0,1.5,2.0,3.0
365 : \endplumedfile
366 : The number of elements in the RECT array should be equal to the number of replicas.
367 :
368 : */
369 : //+ENDPLUMEDOC
370 :
371 : class MetaD : public Bias {
372 :
373 : private:
374 : struct Gaussian {
375 : bool multivariate; // this is required to discriminate the one dimensional case
376 : double height;
377 : std::vector<double> center;
378 : std::vector<double> sigma;
379 : std::vector<double> invsigma;
380 5275 : Gaussian(const bool m, const double h, const std::vector<double>& c, const std::vector<double>& s):
381 5275 : multivariate(m),height(h),center(c),sigma(s),invsigma(s) {
382 : // to avoid troubles from zero element in flexible hills
383 15362 : for(unsigned i=0; i<invsigma.size(); ++i) {
384 10087 : if(std::abs(invsigma[i])>1.e-20) invsigma[i]=1.0/invsigma[i] ;
385 0 : else invsigma[i]=0.0;
386 : }
387 5275 : }
388 : };
389 8 : struct TemperingSpecs {
390 : bool is_active;
391 : std::string name_stem;
392 : std::string name;
393 : double biasf;
394 : double threshold;
395 : double alpha;
396 156 : inline TemperingSpecs(bool is_active, const std::string &name_stem, const std::string &name, double biasf, double threshold, double alpha) :
397 156 : is_active(is_active), name_stem(name_stem), name(name), biasf(biasf), threshold(threshold), alpha(alpha)
398 156 : {}
399 : };
400 : // general setup
401 : double kbt_;
402 : int stride_;
403 : bool calc_work_;
404 : // well-tempered MetaD
405 : bool welltemp_;
406 : double biasf_;
407 : // output files format
408 : std::string fmt_;
409 : // first step?
410 : bool isFirstStep_;
411 : // Gaussian starting parameters
412 : double height0_;
413 : std::vector<double> sigma0_;
414 : std::vector<double> sigma0min_;
415 : std::vector<double> sigma0max_;
416 : // Gaussians
417 : std::vector<Gaussian> hills_;
418 : std::unique_ptr<FlexibleBin> flexbin_;
419 : int adaptive_;
420 : OFile hillsOfile_;
421 : std::vector<std::unique_ptr<IFile>> ifiles_;
422 : std::vector<std::string> ifilesnames_;
423 : // Grids
424 : bool grid_;
425 : std::unique_ptr<GridBase> BiasGrid_;
426 : OFile gridfile_;
427 : bool storeOldGrids_;
428 : int wgridstride_;
429 : // multiple walkers
430 : int mw_n_;
431 : std::string mw_dir_;
432 : int mw_id_;
433 : int mw_rstride_;
434 : bool walkers_mpi_;
435 : unsigned mpi_nw_;
436 : // flying gaussians
437 : bool flying_;
438 : // kinetics from metadynamics
439 : bool acceleration_;
440 : double acc_;
441 : double acc_restart_mean_;
442 : // transition-tempering metadynamics
443 : bool calc_max_bias_;
444 : double max_bias_;
445 : bool calc_transition_bias_;
446 : double transition_bias_;
447 : std::vector<std::vector<double> > transitionwells_;
448 : static const size_t n_tempering_options_ = 1;
449 : static const std::string tempering_names_[1][2];
450 : double dampfactor_;
451 : struct TemperingSpecs tt_specs_;
452 : std::string targetfilename_;
453 : std::unique_ptr<GridBase> TargetGrid_;
454 : // frequency adaptive metadynamics
455 : int current_stride_;
456 : bool freq_adaptive_;
457 : int fa_update_frequency_;
458 : int fa_max_stride_;
459 : double fa_min_acceleration_;
460 : // intervals
461 : double uppI_;
462 : double lowI_;
463 : bool doInt_;
464 : // reweighting
465 : bool calc_rct_;
466 : double reweight_factor_;
467 : unsigned rct_ustride_;
468 : // work
469 : double work_;
470 : // neighbour list stuff
471 : bool nlist_;
472 : bool nlist_update_;
473 : unsigned nlist_steps_;
474 : std::array<double,2> nlist_param_;
475 : std::vector<Gaussian> nlist_hills_;
476 : std::vector<double> nlist_center_;
477 : std::vector<double> nlist_dev2_;
478 :
479 : double stretchA=1.0;
480 : double stretchB=0.0;
481 :
482 : bool noStretchWarningDone=false;
483 :
484 12 : void noStretchWarning() {
485 12 : if(!noStretchWarningDone) {
486 3 : log<<"\nWARNING: you are using a HILLS file with Gaussian kernels, PLUMED 2.8 uses stretched Gaussians by default\n";
487 : }
488 12 : noStretchWarningDone=true;
489 12 : }
490 :
491 : static void registerTemperingKeywords(const std::string &name_stem, const std::string &name, Keywords &keys);
492 : void readTemperingSpecs(TemperingSpecs &t_specs);
493 : void logTemperingSpecs(const TemperingSpecs &t_specs);
494 : void readGaussians(IFile*);
495 : void writeGaussian(const Gaussian&,OFile&);
496 : void addGaussian(const Gaussian&);
497 : double getHeight(const std::vector<double>&);
498 : void temperHeight(double &height, const TemperingSpecs &t_specs, const double tempering_bias);
499 : double getBias(const std::vector<double>&);
500 : double getBiasAndDerivatives(const std::vector<double>&, std::vector<double>&);
501 : double evaluateGaussian(const std::vector<double>&, const Gaussian&);
502 : double evaluateGaussianAndDerivatives(const std::vector<double>&, const Gaussian&,std::vector<double>&,std::vector<double>&);
503 : double getGaussianNormalization(const Gaussian&);
504 : std::vector<unsigned> getGaussianSupport(const Gaussian&);
505 : bool scanOneHill(IFile* ifile, std::vector<Value>& v, std::vector<double>& center, std::vector<double>& sigma, double& height, bool& multivariate);
506 : void computeReweightingFactor();
507 : double getTransitionBarrierBias();
508 : void updateFrequencyAdaptiveStride();
509 : void updateNlist();
510 :
511 : public:
512 : explicit MetaD(const ActionOptions&);
513 : void calculate() override;
514 : void update() override;
515 : static void registerKeywords(Keywords& keys);
516 : bool checkNeedsGradients()const override;
517 : };
518 :
519 : PLUMED_REGISTER_ACTION(MetaD,"METAD")
520 :
521 159 : void MetaD::registerKeywords(Keywords& keys) {
522 159 : Bias::registerKeywords(keys);
523 318 : keys.addOutputComponent("rbias","CALC_RCT","the instantaneous value of the bias normalized using the c(t) reweighting factor [rbias=bias-rct]."
524 : "This component can be used to obtain a reweighted histogram.");
525 318 : keys.addOutputComponent("rct","CALC_RCT","the reweighting factor c(t).");
526 318 : keys.addOutputComponent("work","CALC_WORK","accumulator for work");
527 318 : keys.addOutputComponent("acc","ACCELERATION","the metadynamics acceleration factor");
528 318 : keys.addOutputComponent("maxbias", "CALC_MAX_BIAS", "the maximum of the metadynamics V(s, t)");
529 318 : keys.addOutputComponent("transbias", "CALC_TRANSITION_BIAS", "the metadynamics transition bias V*(t)");
530 318 : keys.addOutputComponent("pace","FREQUENCY_ADAPTIVE","the hill addition frequency when employing frequency adaptive metadynamics");
531 318 : keys.addOutputComponent("nlker","NLIST","number of hills in the neighbor list");
532 318 : keys.addOutputComponent("nlsteps","NLIST","number of steps from last neighbor list update");
533 159 : keys.use("ARG");
534 318 : keys.add("compulsory","SIGMA","the widths of the Gaussian hills");
535 318 : keys.add("compulsory","PACE","the frequency for hill addition");
536 318 : keys.add("compulsory","FILE","HILLS","a file in which the list of added hills is stored");
537 318 : keys.add("optional","HEIGHT","the heights of the Gaussian hills. Compulsory unless TAU and either BIASFACTOR or DAMPFACTOR are given");
538 318 : keys.add("optional","FMT","specify format for HILLS files (useful for decrease the number of digits in regtests)");
539 318 : keys.add("optional","BIASFACTOR","use well tempered metadynamics and use this bias factor. Please note you must also specify temp");
540 318 : keys.addFlag("CALC_WORK",false,"calculate the total accumulated work done by the bias since last restart");
541 318 : keys.add("optional","RECT","list of bias factors for all the replicas");
542 318 : keys.add("optional","DAMPFACTOR","damp hills with exp(-max(V)/(kT*DAMPFACTOR)");
543 318 : for (size_t i = 0; i < n_tempering_options_; i++) {
544 159 : registerTemperingKeywords(tempering_names_[i][0], tempering_names_[i][1], keys);
545 : }
546 318 : keys.add("optional","TARGET","target to a predefined distribution");
547 318 : keys.add("optional","TEMP","the system temperature - this is only needed if you are doing well-tempered metadynamics");
548 318 : keys.add("optional","TAU","in well tempered metadynamics, sets height to (k_B Delta T*pace*timestep)/tau");
549 318 : keys.addFlag("CALC_RCT",false,"calculate the c(t) reweighting factor and use that to obtain the normalized bias [rbias=bias-rct]."
550 : "This method is not compatible with metadynamics not on a grid.");
551 318 : keys.add("optional","RCT_USTRIDE","the update stride for calculating the c(t) reweighting factor."
552 : "The default 1, so c(t) is updated every time the bias is updated.");
553 318 : keys.add("optional","GRID_MIN","the lower bounds for the grid");
554 318 : keys.add("optional","GRID_MAX","the upper bounds for the grid");
555 318 : keys.add("optional","GRID_BIN","the number of bins for the grid");
556 318 : keys.add("optional","GRID_SPACING","the approximate grid spacing (to be used as an alternative or together with GRID_BIN)");
557 318 : keys.addFlag("GRID_SPARSE",false,"use a sparse grid to store hills");
558 318 : keys.addFlag("GRID_NOSPLINE",false,"don't use spline interpolation with grids");
559 318 : keys.add("optional","GRID_WSTRIDE","write the grid to a file every N steps");
560 318 : keys.add("optional","GRID_WFILE","the file on which to write the grid");
561 318 : keys.add("optional","GRID_RFILE","a grid file from which the bias should be read at the initial step of the simulation");
562 318 : keys.addFlag("STORE_GRIDS",false,"store all the grid files the calculation generates. They will be deleted if this keyword is not present");
563 318 : keys.addFlag("NLIST",false,"Use neighbor list for kernels summation, faster but experimental");
564 318 : keys.add("optional", "NLIST_PARAMETERS","(default=6.,0.5) the two cutoff parameters for the Gaussians neighbor list");
565 318 : keys.add("optional","ADAPTIVE","use a geometric (=GEOM) or diffusion (=DIFF) based hills width scheme. Sigma is one number that has distance units or time step dimensions");
566 318 : keys.add("optional","SIGMA_MAX","the upper bounds for the sigmas (in CV units) when using adaptive hills. Negative number means no bounds ");
567 318 : keys.add("optional","SIGMA_MIN","the lower bounds for the sigmas (in CV units) when using adaptive hills. Negative number means no bounds ");
568 318 : keys.add("optional","WALKERS_ID", "walker id");
569 318 : keys.add("optional","WALKERS_N", "number of walkers");
570 318 : keys.add("optional","WALKERS_DIR", "shared directory with the hills files from all the walkers");
571 318 : keys.add("optional","WALKERS_RSTRIDE","stride for reading hills files");
572 318 : keys.addFlag("WALKERS_MPI",false,"Switch on MPI version of multiple walkers - not compatible with WALKERS_* options other than WALKERS_DIR");
573 318 : keys.add("optional","INTERVAL","one dimensional lower and upper limits, outside the limits the system will not feel the biasing force.");
574 318 : keys.addFlag("FLYING_GAUSSIAN",false,"Switch on flying Gaussian method, must be used with WALKERS_MPI");
575 318 : keys.addFlag("ACCELERATION",false,"Set to TRUE if you want to compute the metadynamics acceleration factor.");
576 318 : keys.add("optional","ACCELERATION_RFILE","a data file from which the acceleration should be read at the initial step of the simulation");
577 318 : keys.addFlag("CALC_MAX_BIAS", false, "Set to TRUE if you want to compute the maximum of the metadynamics V(s, t)");
578 318 : keys.addFlag("CALC_TRANSITION_BIAS", false, "Set to TRUE if you want to compute a metadynamics transition bias V*(t)");
579 318 : keys.add("numbered", "TRANSITIONWELL", "This keyword appears multiple times as TRANSITIONWELL followed by an integer. Each specifies the coordinates for one well as in transition-tempered metadynamics. At least one must be provided.");
580 318 : keys.addFlag("FREQUENCY_ADAPTIVE",false,"Set to TRUE if you want to enable frequency adaptive metadynamics such that the frequency for hill addition to change dynamically based on the acceleration factor.");
581 318 : keys.add("optional","FA_UPDATE_FREQUENCY","the frequency for updating the hill addition pace in frequency adaptive metadynamics, by default this is equal to the value given in PACE");
582 318 : keys.add("optional","FA_MAX_PACE","the maximum hill addition frequency allowed in frequency adaptive metadynamics. By default there is no maximum value.");
583 318 : keys.add("optional","FA_MIN_ACCELERATION","only update the hill addition pace in frequency adaptive metadynamics after reaching the minimum acceleration factor given here. By default it is 1.0.");
584 159 : keys.use("RESTART");
585 159 : keys.use("UPDATE_FROM");
586 159 : keys.use("UPDATE_UNTIL");
587 159 : }
588 :
589 : const std::string MetaD::tempering_names_[1][2] = {{"TT", "transition tempered"}};
590 :
591 159 : void MetaD::registerTemperingKeywords(const std::string &name_stem, const std::string &name, Keywords &keys) {
592 318 : keys.add("optional", name_stem + "BIASFACTOR", "use " + name + " metadynamics with this bias factor. Please note you must also specify temp");
593 318 : keys.add("optional", name_stem + "BIASTHRESHOLD", "use " + name + " metadynamics with this bias threshold. Please note you must also specify " + name_stem + "BIASFACTOR");
594 318 : keys.add("optional", name_stem + "ALPHA", "use " + name + " metadynamics with this hill size decay exponent parameter. Please note you must also specify " + name_stem + "BIASFACTOR");
595 159 : }
596 :
597 157 : MetaD::MetaD(const ActionOptions& ao):
598 : PLUMED_BIAS_INIT(ao),
599 156 : kbt_(0.0),
600 156 : stride_(0),
601 156 : calc_work_(false),
602 156 : welltemp_(false),
603 156 : biasf_(-1.0),
604 156 : isFirstStep_(true),
605 156 : height0_(std::numeric_limits<double>::max()),
606 156 : adaptive_(FlexibleBin::none),
607 156 : grid_(false),
608 156 : wgridstride_(0),
609 156 : mw_n_(1), mw_dir_(""), mw_id_(0), mw_rstride_(1),
610 156 : walkers_mpi_(false), mpi_nw_(0),
611 156 : flying_(false),
612 156 : acceleration_(false), acc_(0.0), acc_restart_mean_(0.0),
613 156 : calc_max_bias_(false), max_bias_(0.0),
614 156 : calc_transition_bias_(false), transition_bias_(0.0),
615 156 : dampfactor_(0.0),
616 312 : tt_specs_(false, "TT", "Transition Tempered", -1.0, 0.0, 1.0),
617 156 : current_stride_(0),
618 156 : freq_adaptive_(false),
619 156 : fa_update_frequency_(0),
620 156 : fa_max_stride_(0),
621 156 : fa_min_acceleration_(1.0),
622 156 : uppI_(-1), lowI_(-1), doInt_(false),
623 156 : calc_rct_(false),
624 156 : reweight_factor_(0.0),
625 156 : rct_ustride_(1),
626 156 : work_(0),
627 156 : nlist_(false),
628 156 : nlist_update_(false),
629 469 : nlist_steps_(0)
630 : {
631 156 : if(!dp2cutoffNoStretch()) {
632 156 : stretchA=dp2cutoffA;
633 156 : stretchB=dp2cutoffB;
634 : }
635 : // parse the flexible hills
636 : std::string adaptiveoption;
637 : adaptiveoption="NONE";
638 312 : parse("ADAPTIVE",adaptiveoption);
639 156 : if(adaptiveoption=="GEOM") {
640 22 : log.printf(" Uses Geometry-based hills width: sigma must be in distance units and only one sigma is needed\n");
641 22 : adaptive_=FlexibleBin::geometry;
642 134 : } else if(adaptiveoption=="DIFF") {
643 3 : log.printf(" Uses Diffusion-based hills width: sigma must be in time steps and only one sigma is needed\n");
644 3 : adaptive_=FlexibleBin::diffusion;
645 131 : } else if(adaptiveoption=="NONE") {
646 130 : adaptive_=FlexibleBin::none;
647 : } else {
648 1 : error("I do not know this type of adaptive scheme");
649 : }
650 :
651 155 : parse("FMT",fmt_);
652 :
653 : // parse the sigma
654 155 : parseVector("SIGMA",sigma0_);
655 155 : if(adaptive_==FlexibleBin::none) {
656 : // if you use normal sigma you need one sigma per argument
657 130 : if( sigma0_.size()!=getNumberOfArguments() ) error("number of arguments does not match number of SIGMA parameters");
658 : } else {
659 : // if you use flexible hills you need one sigma
660 25 : if(sigma0_.size()!=1) {
661 1 : error("If you choose ADAPTIVE you need only one sigma according to your choice of type (GEOM/DIFF)");
662 : }
663 : // if adaptive then the number must be an integer
664 24 : if(adaptive_==FlexibleBin::diffusion) {
665 3 : if(int(sigma0_[0])-sigma0_[0]>1.e-9 || int(sigma0_[0])-sigma0_[0] <-1.e-9 || int(sigma0_[0])<1 ) {
666 0 : error("In case of adaptive hills with diffusion, the sigma must be an integer which is the number of time steps\n");
667 : }
668 : }
669 : // here evtl parse the sigma min and max values
670 48 : parseVector("SIGMA_MIN",sigma0min_);
671 24 : if(sigma0min_.size()>0 && sigma0min_.size()!=getNumberOfArguments()) {
672 1 : error("the number of SIGMA_MIN values be the same of the number of the arguments");
673 23 : } else if(sigma0min_.size()==0) {
674 23 : sigma0min_.resize(getNumberOfArguments());
675 67 : for(unsigned i=0; i<getNumberOfArguments(); i++) {sigma0min_[i]=-1.;}
676 : }
677 :
678 46 : parseVector("SIGMA_MAX",sigma0max_);
679 23 : if(sigma0max_.size()>0 && sigma0max_.size()!=getNumberOfArguments()) {
680 1 : error("the number of SIGMA_MAX values be the same of the number of the arguments");
681 22 : } else if(sigma0max_.size()==0) {
682 22 : sigma0max_.resize(getNumberOfArguments());
683 64 : for(unsigned i=0; i<getNumberOfArguments(); i++) {sigma0max_[i]=-1.;}
684 : }
685 :
686 44 : flexbin_=Tools::make_unique<FlexibleBin>(adaptive_,this,sigma0_[0],sigma0min_,sigma0max_);
687 : }
688 :
689 : // note: HEIGHT is not compulsory, since one could use the TAU keyword, see below
690 152 : parse("HEIGHT",height0_);
691 152 : parse("PACE",stride_);
692 151 : if(stride_<=0) error("frequency for hill addition is nonsensical");
693 151 : current_stride_ = stride_;
694 159 : std::string hillsfname="HILLS";
695 151 : parse("FILE",hillsfname);
696 :
697 : // Manually set to calculate special bias quantities
698 : // throughout the course of simulation. (These are chosen due to
699 : // relevance for tempering and event-driven logic as well.)
700 151 : parseFlag("CALC_MAX_BIAS", calc_max_bias_);
701 305 : parseFlag("CALC_TRANSITION_BIAS", calc_transition_bias_);
702 :
703 : std::vector<double> rect_biasf_;
704 302 : parseVector("RECT",rect_biasf_);
705 151 : if(rect_biasf_.size()>0) {
706 18 : int r=0;
707 18 : if(comm.Get_rank()==0) r=multi_sim_comm.Get_rank();
708 18 : comm.Bcast(r,0);
709 18 : biasf_=rect_biasf_[r];
710 18 : log<<" You are using RECT\n";
711 : } else {
712 266 : parse("BIASFACTOR",biasf_);
713 : }
714 151 : if( biasf_<1.0 && biasf_!=-1.0) error("well tempered bias factor is nonsensical");
715 302 : parse("DAMPFACTOR",dampfactor_); kbt_=getkBT();
716 151 : if(biasf_>=1.0) {
717 38 : if(kbt_==0.0) error("Unless the MD engine passes the temperature to plumed, with well-tempered metad you must specify it using TEMP");
718 38 : welltemp_=true;
719 : }
720 151 : if(dampfactor_>0.0) {
721 2 : if(kbt_==0.0) error("Unless the MD engine passes the temperature to plumed, with damped metad you must specify it using TEMP");
722 : }
723 :
724 151 : parseFlag("CALC_WORK",calc_work_);
725 :
726 : // Set transition tempering parameters.
727 : // Transition wells are read later via calc_transition_bias_.
728 151 : readTemperingSpecs(tt_specs_);
729 151 : if (tt_specs_.is_active) calc_transition_bias_ = true;
730 :
731 : // If any previous option specified to calculate a transition bias,
732 : // now read the transition wells for that quantity.
733 151 : if (calc_transition_bias_) {
734 13 : std::vector<double> tempcoords(getNumberOfArguments());
735 26 : for (unsigned i = 0; ; i++) {
736 78 : if (!parseNumberedVector("TRANSITIONWELL", i, tempcoords) ) break;
737 26 : if (tempcoords.size() != getNumberOfArguments()) {
738 0 : error("incorrect number of coordinates for transition tempering well");
739 : }
740 26 : transitionwells_.push_back(tempcoords);
741 : }
742 : }
743 :
744 302 : parse("TARGET",targetfilename_);
745 151 : if(targetfilename_.length()>0 && kbt_==0.0) error("with TARGET temperature must be specified");
746 151 : double tau=0.0;
747 151 : parse("TAU",tau);
748 151 : if(tau==0.0) {
749 129 : if(height0_==std::numeric_limits<double>::max()) error("At least one between HEIGHT and TAU should be specified");
750 : // if tau is not set, we compute it here from the other input parameters
751 129 : if(welltemp_) tau=(kbt_*(biasf_-1.0))/height0_*getTimeStep()*stride_;
752 110 : else if(dampfactor_>0.0) tau=(kbt_*dampfactor_)/height0_*getTimeStep()*stride_;
753 : } else {
754 22 : if(height0_!=std::numeric_limits<double>::max()) error("At most one between HEIGHT and TAU should be specified");
755 22 : if(welltemp_) {
756 19 : if(biasf_!=1.0) height0_=(kbt_*(biasf_-1.0))/tau*getTimeStep()*stride_;
757 4 : else height0_=kbt_/tau*getTimeStep()*stride_; // special case for gamma=1
758 : }
759 3 : else if(dampfactor_>0.0) height0_=(kbt_*dampfactor_)/tau*getTimeStep()*stride_;
760 1 : else error("TAU only makes sense in well-tempered or damped metadynamics");
761 : }
762 :
763 : // Grid Stuff
764 153 : std::vector<std::string> gmin(getNumberOfArguments());
765 300 : parseVector("GRID_MIN",gmin);
766 150 : if(gmin.size()!=getNumberOfArguments() && gmin.size()!=0) error("not enough values for GRID_MIN");
767 150 : std::vector<std::string> gmax(getNumberOfArguments());
768 300 : parseVector("GRID_MAX",gmax);
769 150 : if(gmax.size()!=getNumberOfArguments() && gmax.size()!=0) error("not enough values for GRID_MAX");
770 150 : std::vector<unsigned> gbin(getNumberOfArguments());
771 : std::vector<double> gspacing;
772 300 : parseVector("GRID_BIN",gbin);
773 150 : if(gbin.size()!=getNumberOfArguments() && gbin.size()!=0) error("not enough values for GRID_BIN");
774 300 : parseVector("GRID_SPACING",gspacing);
775 150 : if(gspacing.size()!=getNumberOfArguments() && gspacing.size()!=0) error("not enough values for GRID_SPACING");
776 150 : if(gmin.size()!=gmax.size()) error("GRID_MAX and GRID_MIN should be either present or absent");
777 150 : if(gspacing.size()!=0 && gmin.size()==0) error("If GRID_SPACING is present also GRID_MIN and GRID_MAX should be present");
778 150 : if(gbin.size()!=0 && gmin.size()==0) error("If GRID_BIN is present also GRID_MIN and GRID_MAX should be present");
779 150 : if(gmin.size()!=0) {
780 61 : if(gbin.size()==0 && gspacing.size()==0) {
781 6 : if(adaptive_==FlexibleBin::none) {
782 6 : log<<" Binsize not specified, 1/5 of sigma will be be used\n";
783 6 : plumed_assert(sigma0_.size()==getNumberOfArguments());
784 6 : gspacing.resize(getNumberOfArguments());
785 13 : for(unsigned i=0; i<gspacing.size(); i++) gspacing[i]=0.2*sigma0_[i];
786 : } else {
787 : // with adaptive hills and grid a sigma min must be specified
788 0 : for(unsigned i=0; i<sigma0min_.size(); i++) if(sigma0min_[i]<=0) error("When using ADAPTIVE Gaussians on a grid SIGMA_MIN must be specified");
789 0 : log<<" Binsize not specified, 1/5 of sigma_min will be be used\n";
790 0 : gspacing.resize(getNumberOfArguments());
791 0 : for(unsigned i=0; i<gspacing.size(); i++) gspacing[i]=0.2*sigma0min_[i];
792 : }
793 55 : } else if(gspacing.size()!=0 && gbin.size()==0) {
794 2 : log<<" The number of bins will be estimated from GRID_SPACING\n";
795 53 : } else if(gspacing.size()!=0 && gbin.size()!=0) {
796 1 : log<<" You specified both GRID_BIN and GRID_SPACING\n";
797 1 : log<<" The more conservative (highest) number of bins will be used for each variable\n";
798 : }
799 61 : if(gbin.size()==0) gbin.assign(getNumberOfArguments(),1);
800 73 : if(gspacing.size()!=0) for(unsigned i=0; i<getNumberOfArguments(); i++) {
801 : double a,b;
802 13 : Tools::convert(gmin[i],a);
803 12 : Tools::convert(gmax[i],b);
804 12 : unsigned n=std::ceil(((b-a)/gspacing[i]));
805 12 : if(gbin[i]<n) gbin[i]=n;
806 : }
807 : }
808 149 : if(gbin.size()>0) grid_=true;
809 :
810 149 : bool sparsegrid=false;
811 149 : parseFlag("GRID_SPARSE",sparsegrid);
812 149 : bool nospline=false;
813 149 : parseFlag("GRID_NOSPLINE",nospline);
814 149 : bool spline=!nospline;
815 300 : parse("GRID_WSTRIDE",wgridstride_);
816 : std::string gridfilename_;
817 149 : parse("GRID_WFILE",gridfilename_);
818 149 : parseFlag("STORE_GRIDS",storeOldGrids_);
819 149 : if(grid_ && gridfilename_.length()>0) {
820 19 : if(wgridstride_==0 ) error("frequency with which to output grid not specified use GRID_WSTRIDE");
821 : }
822 149 : if(grid_ && wgridstride_>0) {
823 20 : if(gridfilename_.length()==0) error("grid filename not specified use GRID_WFILE");
824 : }
825 :
826 : std::string gridreadfilename_;
827 149 : parse("GRID_RFILE",gridreadfilename_);
828 :
829 149 : if(!grid_&&gridfilename_.length()> 0) error("To write a grid you need first to define it!");
830 149 : if(!grid_&&gridreadfilename_.length()>0) error("To read a grid you need first to define it!");
831 :
832 : /*setup neighbor list stuff*/
833 298 : parseFlag("NLIST", nlist_);
834 149 : nlist_center_.resize(getNumberOfArguments());
835 149 : nlist_dev2_.resize(getNumberOfArguments());
836 150 : if(nlist_&&grid_) error("NLIST and GRID cannot be combined!");
837 : std::vector<double> nlist_param;
838 298 : parseVector("NLIST_PARAMETERS",nlist_param);
839 149 : if(nlist_param.size()==0)
840 : {
841 149 : nlist_param_[0]=6.0;//*DP2CUTOFF -> max distance of neighbors
842 149 : nlist_param_[1]=0.5;//*nlist_dev2_[i] -> condition for rebuilding
843 : }
844 : else
845 : {
846 0 : plumed_massert(nlist_param.size()==2,"two cutoff parameters are needed for the neighbor list");
847 0 : plumed_massert(nlist_param[0]>1.0,"the first of NLIST_PARAMETERS must be greater than 1. The smaller the first, the smaller should be the second as well");
848 0 : const double min_PARAM_1=(1.-1./std::sqrt(nlist_param[0]/2))+0.16;
849 0 : plumed_massert(nlist_param[1]>0,"the second of NLIST_PARAMETERS must be greater than 0");
850 0 : plumed_massert(nlist_param[1]<=min_PARAM_1,"the second of NLIST_PARAMETERS must be smaller to avoid systematic errors. Largest suggested value is: 1.16-1/sqrt(PARAM_0/2) = "+std::to_string(min_PARAM_1));
851 0 : nlist_param_[0]=nlist_param[0];
852 0 : nlist_param_[1]=nlist_param[1];
853 : }
854 :
855 : // Reweighting factor rct
856 149 : parseFlag("CALC_RCT",calc_rct_);
857 149 : if (calc_rct_) plumed_massert(grid_,"CALC_RCT is supported only if bias is on a grid");
858 149 : parse("RCT_USTRIDE",rct_ustride_);
859 :
860 149 : if(dampfactor_>0.0) {
861 2 : if(!grid_) error("With DAMPFACTOR you should use grids");
862 : }
863 :
864 : // Multiple walkers
865 149 : parse("WALKERS_N",mw_n_);
866 149 : parse("WALKERS_ID",mw_id_);
867 149 : if(mw_n_<=mw_id_) error("walker ID should be a numerical value less than the total number of walkers");
868 149 : parse("WALKERS_DIR",mw_dir_);
869 149 : parse("WALKERS_RSTRIDE",mw_rstride_);
870 :
871 : // MPI version
872 149 : parseFlag("WALKERS_MPI",walkers_mpi_);
873 :
874 : //If this Action is not compiled with MPI the user is informed and we exit gracefully
875 149 : if(walkers_mpi_) {
876 39 : plumed_assert(Communicator::plumedHasMPI()) << "Invalid walkers configuration: WALKERS_MPI flag requires MPI compilation";
877 40 : plumed_assert(Communicator::initialized()) << "Invalid walkers configuration: WALKERS_MPI needs the communicator correctly initialized.";
878 : }
879 :
880 : // Flying Gaussian
881 148 : parseFlag("FLYING_GAUSSIAN", flying_);
882 :
883 : // Inteval keyword
884 149 : std::vector<double> tmpI(2);
885 296 : parseVector("INTERVAL",tmpI);
886 148 : if(tmpI.size()!=2&&tmpI.size()!=0) error("both a lower and an upper limits must be provided with INTERVAL");
887 148 : else if(tmpI.size()==2) {
888 2 : lowI_=tmpI.at(0);
889 2 : uppI_=tmpI.at(1);
890 2 : if(getNumberOfArguments()!=1) error("INTERVAL limits correction works only for monodimensional metadynamics!");
891 2 : if(uppI_<lowI_) error("The Upper limit must be greater than the Lower limit!");
892 2 : if(getPntrToArgument(0)->isPeriodic()) error("INTERVAL cannot be used with periodic variables!");
893 2 : doInt_=true;
894 : }
895 :
896 296 : parseFlag("ACCELERATION",acceleration_);
897 : // Check for a restart acceleration if acceleration is active.
898 : std::string acc_rfilename;
899 148 : if(acceleration_) {
900 8 : parse("ACCELERATION_RFILE", acc_rfilename);
901 : }
902 :
903 148 : freq_adaptive_=false;
904 148 : parseFlag("FREQUENCY_ADAPTIVE",freq_adaptive_);
905 : //
906 148 : fa_update_frequency_=0;
907 148 : parse("FA_UPDATE_FREQUENCY",fa_update_frequency_);
908 148 : if(fa_update_frequency_!=0 && !freq_adaptive_) {
909 0 : plumed_merror("It doesn't make sense to use the FA_MAX_PACE keyword if frequency adaptive METAD hasn't been activated by using the FREQUENCY_ADAPTIVE flag");
910 : }
911 148 : if(fa_update_frequency_==0 && freq_adaptive_) {
912 0 : fa_update_frequency_=stride_;
913 : }
914 : //
915 148 : fa_max_stride_=0;
916 148 : parse("FA_MAX_PACE",fa_max_stride_);
917 148 : if(fa_max_stride_!=0 && !freq_adaptive_) {
918 0 : plumed_merror("It doesn't make sense to use the FA_MAX_PACE keyword if frequency adaptive METAD hasn't been activated by using the FREQUENCY_ADAPTIVE flag");
919 : }
920 : //
921 148 : fa_min_acceleration_=1.0;
922 148 : parse("FA_MIN_ACCELERATION",fa_min_acceleration_);
923 148 : if(fa_min_acceleration_!=1.0 && !freq_adaptive_) {
924 0 : plumed_merror("It doesn't make sense to use the FA_MIN_ACCELERATION keyword if frequency adaptive METAD hasn't been activated by using the FREQUENCY_ADAPTIVE flag");
925 : }
926 :
927 148 : checkRead();
928 :
929 148 : log.printf(" Gaussian width ");
930 148 : if (adaptive_==FlexibleBin::diffusion)log.printf(" (Note: The units of sigma are in timesteps) ");
931 148 : if (adaptive_==FlexibleBin::geometry)log.printf(" (Note: The units of sigma are in dist units) ");
932 396 : for(unsigned i=0; i<sigma0_.size(); ++i) log.printf(" %f",sigma0_[i]);
933 148 : log.printf(" Gaussian height %f\n",height0_);
934 148 : log.printf(" Gaussian deposition pace %d\n",stride_);
935 148 : log.printf(" Gaussian file %s\n",hillsfname.c_str());
936 148 : if(welltemp_) {
937 38 : log.printf(" Well-Tempered Bias Factor %f\n",biasf_);
938 38 : log.printf(" Hills relaxation time (tau) %f\n",tau);
939 38 : log.printf(" KbT %f\n",kbt_);
940 : }
941 :
942 : // Transition tempered metadynamics options
943 148 : if (tt_specs_.is_active) {
944 3 : logTemperingSpecs(tt_specs_);
945 : // Check that the appropriate transition bias quantity is calculated.
946 : // (Should never trip, given that the flag is automatically set.)
947 3 : if (!calc_transition_bias_) {
948 0 : error(" transition tempering requires calculation of a transition bias");
949 : }
950 : }
951 :
952 : // Overall tempering sanity check (this gets tricky when multiple are active).
953 : // When multiple temperings are active, it's fine to have one tempering attempt
954 : // to increase hill size with increasing bias, so long as the others can shrink
955 : // the hills faster than it increases their size in the long-time limit.
956 : // This set of checks ensures that the hill sizes eventually decay to zero as c(t)
957 : // diverges to infinity.
958 : // The alpha parameter allows hills to decay as 1/t^alpha instead of 1/t,
959 : // a slower decay, so as t -> infinity, only the temperings with the largest
960 : // alphas govern the final asymptotic decay. (Alpha helps prevent false convergence.)
961 148 : if (welltemp_ || dampfactor_ > 0.0 || tt_specs_.is_active) {
962 : // Determine the number of active temperings.
963 : int n_active = 0;
964 43 : if (welltemp_) n_active++;
965 43 : if (dampfactor_ > 0.0) n_active++;
966 43 : if (tt_specs_.is_active) n_active++;
967 : // Find the greatest alpha.
968 43 : double greatest_alpha = 0.0;
969 43 : if (welltemp_) greatest_alpha = std::max(greatest_alpha, 1.0);
970 45 : if (dampfactor_ > 0.0) greatest_alpha = std::max(greatest_alpha, 1.0);
971 46 : if (tt_specs_.is_active) greatest_alpha = std::max(greatest_alpha, tt_specs_.alpha);
972 : // Find the least alpha.
973 43 : double least_alpha = 1.0;
974 : if (welltemp_) least_alpha = std::min(least_alpha, 1.0);
975 43 : if (dampfactor_ > 0.0) least_alpha = std::min(least_alpha, 1.0);
976 44 : if (tt_specs_.is_active) least_alpha = std::min(least_alpha, tt_specs_.alpha);
977 : // Find the inverse harmonic average of the delta T parameters for all
978 : // of the temperings with the greatest alpha values.
979 : double total_governing_deltaT_inv = 0.0;
980 43 : if (welltemp_ && 1.0 == greatest_alpha && biasf_ != 1.0) total_governing_deltaT_inv += 1.0 / (biasf_ - 1.0);
981 43 : if (dampfactor_ > 0.0 && 1.0 == greatest_alpha) total_governing_deltaT_inv += 1.0 / (dampfactor_);
982 43 : if (tt_specs_.is_active && tt_specs_.alpha == greatest_alpha) total_governing_deltaT_inv += 1.0 / (tt_specs_.biasf - 1.0);
983 : // Give a newbie-friendly error message for people using one tempering if
984 : // only one is active.
985 43 : if (n_active == 1 && total_governing_deltaT_inv < 0.0) {
986 0 : error("for stable tempering, the bias factor must be greater than one");
987 : // Give a slightly more complex error message to users stacking multiple
988 : // tempering options at a time, but all with uniform alpha values.
989 43 : } else if (total_governing_deltaT_inv < 0.0 && greatest_alpha == least_alpha) {
990 0 : error("for stable tempering, the sum of the inverse Delta T parameters must be greater than zero!");
991 : // Give the most technical error message to users stacking multiple tempering
992 : // options with different alpha parameters.
993 43 : } else if (total_governing_deltaT_inv < 0.0 && greatest_alpha != least_alpha) {
994 0 : error("for stable tempering, the sum of the inverse Delta T parameters for the greatest asymptotic hill decay exponents must be greater than zero!");
995 : }
996 : }
997 :
998 148 : if(doInt_) log.printf(" Upper and Lower limits boundaries for the bias are activated at %f - %f\n", lowI_, uppI_);
999 :
1000 148 : if(grid_) {
1001 60 : log.printf(" Grid min");
1002 161 : for(unsigned i=0; i<gmin.size(); ++i) log.printf(" %s",gmin[i].c_str() );
1003 60 : log.printf("\n");
1004 60 : log.printf(" Grid max");
1005 161 : for(unsigned i=0; i<gmax.size(); ++i) log.printf(" %s",gmax[i].c_str() );
1006 60 : log.printf("\n");
1007 60 : log.printf(" Grid bin");
1008 161 : for(unsigned i=0; i<gbin.size(); ++i) log.printf(" %u",gbin[i]);
1009 60 : log.printf("\n");
1010 60 : if(spline) {log.printf(" Grid uses spline interpolation\n");}
1011 60 : if(sparsegrid) {log.printf(" Grid uses sparse grid\n");}
1012 60 : if(wgridstride_>0) {log.printf(" Grid is written on file %s with stride %d\n",gridfilename_.c_str(),wgridstride_);}
1013 : }
1014 :
1015 148 : if(mw_n_>1) {
1016 6 : if(walkers_mpi_) error("MPI version of multiple walkers is not compatible with filesystem version of multiple walkers");
1017 6 : log.printf(" %d multiple walkers active\n",mw_n_);
1018 6 : log.printf(" walker id %d\n",mw_id_);
1019 6 : log.printf(" reading stride %d\n",mw_rstride_);
1020 6 : if(mw_dir_!="")log.printf(" directory with hills files %s\n",mw_dir_.c_str());
1021 : } else {
1022 142 : if(walkers_mpi_) {
1023 38 : log.printf(" Multiple walkers active using MPI communnication\n");
1024 38 : if(mw_dir_!="")log.printf(" directory with hills files %s\n",mw_dir_.c_str());
1025 38 : if(comm.Get_rank()==0) {
1026 : // Only root of group can communicate with other walkers
1027 23 : mpi_nw_=multi_sim_comm.Get_size();
1028 : }
1029 : // Communicate to the other members of the same group
1030 : // info abount number of walkers and walker index
1031 38 : comm.Bcast(mpi_nw_,0);
1032 : }
1033 : }
1034 :
1035 148 : if(flying_) {
1036 6 : if(!walkers_mpi_) error("Flying Gaussian method must be used with MPI version of multiple walkers");
1037 6 : log.printf(" Flying Gaussian method with %d walkers active\n",mpi_nw_);
1038 : }
1039 :
1040 148 : if(nlist_) {
1041 2 : addComponent("nlker");
1042 2 : componentIsNotPeriodic("nlker");
1043 2 : addComponent("nlsteps");
1044 2 : componentIsNotPeriodic("nlsteps");
1045 : }
1046 :
1047 148 : if(calc_rct_) {
1048 18 : addComponent("rbias"); componentIsNotPeriodic("rbias");
1049 12 : addComponent("rct"); componentIsNotPeriodic("rct");
1050 6 : log.printf(" The c(t) reweighting factor will be calculated every %u hills\n",rct_ustride_);
1051 12 : getPntrToComponent("rct")->set(reweight_factor_);
1052 : }
1053 :
1054 150 : if(calc_work_) { addComponent("work"); componentIsNotPeriodic("work"); }
1055 :
1056 148 : if(acceleration_) {
1057 4 : if (kbt_ == 0.0) {
1058 0 : error("The calculation of the acceleration works only if simulation temperature has been defined");
1059 : }
1060 4 : log.printf(" calculation on the fly of the acceleration factor\n");
1061 12 : addComponent("acc"); componentIsNotPeriodic("acc");
1062 : // Set the initial value of the the acceleration.
1063 : // If this is not a restart, set to 1.0.
1064 4 : if (acc_rfilename.length() == 0) {
1065 4 : getPntrToComponent("acc")->set(1.0);
1066 2 : if(getRestart()) {
1067 1 : log.printf(" WARNING: calculating the acceleration factor in a restarted run without reading in the previous value will most likely lead to incorrect results.\n");
1068 1 : log.printf(" You should use the ACCELERATION_RFILE keyword.\n");
1069 : }
1070 : // Otherwise, read and set the restart value.
1071 : } else {
1072 : // Restart of acceleration does not make sense if the restart timestep is zero.
1073 : //if (getStep() == 0) {
1074 : // error("Restarting calculation of acceleration factors works only if simulation timestep is restarted correctly");
1075 : //}
1076 : // Open the ACCELERATION_RFILE.
1077 2 : IFile acc_rfile;
1078 2 : acc_rfile.link(*this);
1079 2 : if(acc_rfile.FileExist(acc_rfilename)) {
1080 2 : acc_rfile.open(acc_rfilename);
1081 : } else {
1082 0 : error("The ACCELERATION_RFILE file you want to read: " + acc_rfilename + ", cannot be found!");
1083 : }
1084 : // Read the file to find the restart acceleration.
1085 2 : double acc_rmean=0.0;
1086 2 : double acc_rtime=0.0;
1087 : bool found=false;
1088 2 : std::string acclabel = getLabel() + ".acc";
1089 2 : acc_rfile.allowIgnoredFields();
1090 248 : while(acc_rfile.scanField("time", acc_rtime)) {
1091 122 : acc_rfile.scanField(acclabel, acc_rmean);
1092 122 : acc_rfile.scanField();
1093 : found=true;
1094 : }
1095 2 : if(!found) error("The ACCELERATION_RFILE file you want to read: " + acc_rfilename + ", does not contain a time field!");
1096 2 : acc_restart_mean_ = acc_rmean;
1097 : // Set component based on the read values.
1098 2 : getPntrToComponent("acc")->set(acc_rmean);
1099 2 : log.printf(" initial acceleration factor read from file %s: value of %f at time %f\n",acc_rfilename.c_str(),acc_rmean,acc_rtime);
1100 2 : }
1101 : }
1102 :
1103 148 : if (calc_max_bias_) {
1104 0 : if (!grid_) error("Calculating the maximum bias on the fly works only with a grid");
1105 0 : log.printf(" calculation on the fly of the maximum bias max(V(s,t)) \n");
1106 0 : addComponent("maxbias");
1107 0 : componentIsNotPeriodic("maxbias");
1108 : }
1109 :
1110 148 : if (calc_transition_bias_) {
1111 13 : if (!grid_) error("Calculating the transition bias on the fly works only with a grid");
1112 13 : log.printf(" calculation on the fly of the transition bias V*(t)\n");
1113 26 : addComponent("transbias");
1114 13 : componentIsNotPeriodic("transbias");
1115 13 : log<<" Number of transition wells "<<transitionwells_.size()<<"\n";
1116 13 : if (transitionwells_.size() == 0) error("Calculating the transition bias on the fly requires definition of at least one transition well");
1117 : // Check that a grid is in use.
1118 13 : if (!grid_) error(" transition barrier finding requires a grid for the bias");
1119 : // Log the wells and check that they are in the grid.
1120 39 : for (unsigned i = 0; i < transitionwells_.size(); i++) {
1121 : // Log the coordinate.
1122 26 : log.printf(" Transition well %d at coordinate ", i);
1123 64 : for (unsigned j = 0; j < getNumberOfArguments(); j++) log.printf("%f ", transitionwells_[i][j]);
1124 26 : log.printf("\n");
1125 : // Check that the coordinate is in the grid.
1126 64 : for (unsigned j = 0; j < getNumberOfArguments(); j++) {
1127 : double max, min;
1128 38 : Tools::convert(gmin[j], min);
1129 38 : Tools::convert(gmax[j], max);
1130 38 : if (transitionwells_[i][j] < min || transitionwells_[i][j] > max) error(" transition well is not in grid");
1131 : }
1132 : }
1133 : }
1134 :
1135 148 : if(freq_adaptive_) {
1136 2 : if(!acceleration_) {
1137 0 : plumed_merror("Frequency adaptive metadynamics only works if the calculation of the acceleration factor is enabled with the ACCELERATION keyword\n");
1138 : }
1139 2 : if(walkers_mpi_) {
1140 0 : plumed_merror("Combining frequency adaptive metadynamics with MPI multiple walkers is not allowed");
1141 : }
1142 :
1143 2 : log.printf(" Frequency adaptive metadynamics enabled\n");
1144 2 : if(getRestart() && acc_rfilename.length() == 0) {
1145 0 : log.printf(" WARNING: using the frequency adaptive scheme in a restarted run without reading in the previous value of the acceleration factor will most likely lead to incorrect results.\n");
1146 0 : log.printf(" You should use the ACCELERATION_RFILE keyword.\n");
1147 : }
1148 2 : log.printf(" The frequency for hill addition will change dynamically based on the metadynamics acceleration factor\n");
1149 2 : log.printf(" The hill addition frequency will be updated every %d steps\n",fa_update_frequency_);
1150 2 : if(fa_min_acceleration_>1.0) {
1151 2 : log.printf(" The hill addition frequency will only be updated once the metadynamics acceleration factor becomes larger than %.1f \n",fa_min_acceleration_);
1152 : }
1153 2 : if(fa_max_stride_!=0) {
1154 2 : log.printf(" The hill addition frequency will not become larger than %d steps\n",fa_max_stride_);
1155 : }
1156 4 : addComponent("pace"); componentIsNotPeriodic("pace");
1157 2 : updateFrequencyAdaptiveStride();
1158 : }
1159 :
1160 : // initializing and checking grid
1161 : bool restartedFromGrid=false; // restart from grid file
1162 148 : if(grid_) {
1163 60 : if(!(gridreadfilename_.length()>0)) {
1164 : // check for mesh and sigma size
1165 116 : for(unsigned i=0; i<getNumberOfArguments(); i++) {
1166 : double a,b;
1167 74 : Tools::convert(gmin[i],a);
1168 74 : Tools::convert(gmax[i],b);
1169 74 : double mesh=(b-a)/((double)gbin[i]);
1170 74 : if(adaptive_==FlexibleBin::none) {
1171 74 : if(mesh>0.5*sigma0_[i]) log<<" WARNING: Using a METAD with a Grid Spacing larger than half of the Gaussians width (SIGMA) can produce artifacts\n";
1172 : } else {
1173 0 : if(sigma0min_[i]<0.) error("When using ADAPTIVE Gaussians on a grid SIGMA_MIN must be specified");
1174 0 : if(mesh>0.5*sigma0min_[i]) log<<" WARNING: to use a METAD with a GRID and ADAPTIVE you need to set a Grid Spacing lower than half of the Gaussians (SIGMA_MIN) \n";
1175 : }
1176 : }
1177 42 : std::string funcl=getLabel() + ".bias";
1178 42 : if(!sparsegrid) {BiasGrid_=Tools::make_unique<Grid>(funcl,getArguments(),gmin,gmax,gbin,spline,true);}
1179 6 : else {BiasGrid_=Tools::make_unique<SparseGrid>(funcl,getArguments(),gmin,gmax,gbin,spline,true);}
1180 42 : std::vector<std::string> actualmin=BiasGrid_->getMin();
1181 42 : std::vector<std::string> actualmax=BiasGrid_->getMax();
1182 116 : for(unsigned i=0; i<getNumberOfArguments(); i++) {
1183 : std::string is;
1184 74 : Tools::convert(i,is);
1185 74 : if(gmin[i]!=actualmin[i]) error("GRID_MIN["+is+"] must be adjusted to "+actualmin[i]+" to fit periodicity");
1186 74 : if(gmax[i]!=actualmax[i]) error("GRID_MAX["+is+"] must be adjusted to "+actualmax[i]+" to fit periodicity");
1187 : }
1188 42 : } else {
1189 : // read the grid in input, find the keys
1190 18 : if(walkers_mpi_&&gridreadfilename_.at(0)!='/') {
1191 : //if possible the root replica will share its current folder so that all walkers will read the same file
1192 0 : const std::string ret = std::filesystem::current_path();
1193 0 : gridreadfilename_ = "/" + gridreadfilename_;
1194 0 : gridreadfilename_ = ret + gridreadfilename_;
1195 0 : if(comm.Get_rank()==0) multi_sim_comm.Bcast(gridreadfilename_,0);
1196 0 : comm.Bcast(gridreadfilename_,0);
1197 : }
1198 18 : IFile gridfile;
1199 18 : gridfile.link(*this);
1200 18 : if(gridfile.FileExist(gridreadfilename_)) {
1201 18 : gridfile.open(gridreadfilename_);
1202 : } else {
1203 0 : error("The GRID file you want to read: " + gridreadfilename_ + ", cannot be found!");
1204 : }
1205 18 : std::string funcl=getLabel() + ".bias";
1206 36 : BiasGrid_=GridBase::create(funcl, getArguments(), gridfile, gmin, gmax, gbin, sparsegrid, spline, true);
1207 18 : if(BiasGrid_->getDimension()!=getNumberOfArguments()) error("mismatch between dimensionality of input grid and number of arguments");
1208 45 : for(unsigned i=0; i<getNumberOfArguments(); ++i) {
1209 54 : if( getPntrToArgument(i)->isPeriodic()!=BiasGrid_->getIsPeriodic()[i] ) error("periodicity mismatch between arguments and input bias");
1210 : double a, b;
1211 27 : Tools::convert(gmin[i],a);
1212 27 : Tools::convert(gmax[i],b);
1213 27 : double mesh=(b-a)/((double)gbin[i]);
1214 27 : if(mesh>0.5*sigma0_[i]) log<<" WARNING: Using a METAD with a Grid Spacing larger than half of the Gaussians width can produce artifacts\n";
1215 : }
1216 18 : log.printf(" Restarting from %s\n",gridreadfilename_.c_str());
1217 18 : if(getRestart()) restartedFromGrid=true;
1218 18 : }
1219 : }
1220 :
1221 : // if we are restarting from GRID and using WALKERS_MPI we can check that all walkers have actually read the grid
1222 148 : if(getRestart()&&walkers_mpi_) {
1223 9 : std::vector<int> restarted(mpi_nw_,0);
1224 9 : if(comm.Get_rank()==0) multi_sim_comm.Allgather(int(restartedFromGrid), restarted);
1225 9 : comm.Bcast(restarted,0);
1226 : int result = std::accumulate(restarted.begin(),restarted.end(),0);
1227 9 : if(result!=0&&result!=mpi_nw_) error("in this WALKERS_MPI run some replica have restarted from GRID while other do not!");
1228 : }
1229 :
1230 186 : if(walkers_mpi_&&mw_dir_==""&&hillsfname.at(0)!='/') {
1231 : //if possible the root replica will share its current folder so that all walkers will read the same file
1232 76 : const std::string ret = std::filesystem::current_path();
1233 : mw_dir_ = ret;
1234 38 : mw_dir_ = mw_dir_ + "/";
1235 38 : if(comm.Get_rank()==0) multi_sim_comm.Bcast(mw_dir_,0);
1236 38 : comm.Bcast(mw_dir_,0);
1237 : }
1238 :
1239 : // creating std::vector of ifile* for hills reading
1240 : // open all files at the beginning and read Gaussians if restarting
1241 : bool restartedFromHills=false; // restart from hills files
1242 308 : for(int i=0; i<mw_n_; ++i) {
1243 : std::string fname;
1244 160 : if(mw_dir_!="") {
1245 47 : if(mw_n_>1) {
1246 9 : std::stringstream out; out << i;
1247 18 : fname = mw_dir_+"/"+hillsfname+"."+out.str();
1248 47 : } else if(walkers_mpi_) {
1249 76 : fname = mw_dir_+"/"+hillsfname;
1250 : } else {
1251 : fname = hillsfname;
1252 : }
1253 : } else {
1254 113 : if(mw_n_>1) {
1255 9 : std::stringstream out; out << i;
1256 18 : fname = hillsfname+"."+out.str();
1257 9 : } else {
1258 : fname = hillsfname;
1259 : }
1260 : }
1261 160 : ifiles_.emplace_back(Tools::make_unique<IFile>());
1262 : // this is just a shortcut pointer to the last element:
1263 : IFile *ifile = ifiles_.back().get();
1264 160 : ifilesnames_.push_back(fname);
1265 160 : ifile->link(*this);
1266 160 : if(ifile->FileExist(fname)) {
1267 35 : ifile->open(fname);
1268 35 : if(getRestart()&&!restartedFromGrid) {
1269 19 : log.printf(" Restarting from %s:",ifilesnames_[i].c_str());
1270 19 : readGaussians(ifiles_[i].get());
1271 : restartedFromHills=true;
1272 : }
1273 35 : ifiles_[i]->reset(false);
1274 : // close only the walker own hills file for later writing
1275 35 : if(i==mw_id_) ifiles_[i]->close();
1276 : } else {
1277 : // in case a file does not exist and we are restarting, complain that the file was not found
1278 125 : if(getRestart()&&!restartedFromGrid) error("restart file "+fname+" not found");
1279 : }
1280 : }
1281 :
1282 : // if we are restarting from FILE and using WALKERS_MPI we can check that all walkers have actually read the FILE
1283 148 : if(getRestart()&&walkers_mpi_) {
1284 9 : std::vector<int> restarted(mpi_nw_,0);
1285 9 : if(comm.Get_rank()==0) multi_sim_comm.Allgather(int(restartedFromHills), restarted);
1286 9 : comm.Bcast(restarted,0);
1287 : int result = std::accumulate(restarted.begin(),restarted.end(),0);
1288 9 : if(result!=0&&result!=mpi_nw_) error("in this WALKERS_MPI run some replica have restarted from FILE while other do not!");
1289 : }
1290 :
1291 148 : comm.Barrier();
1292 : // this barrier is needed when using walkers_mpi
1293 : // to be sure that all files have been read before
1294 : // backing them up
1295 : // it should not be used when walkers_mpi is false otherwise
1296 : // it would introduce troubles when using replicas without METAD
1297 : // (e.g. in bias exchange with a neutral replica)
1298 : // see issue #168 on github
1299 148 : if(comm.Get_rank()==0 && walkers_mpi_) multi_sim_comm.Barrier();
1300 :
1301 148 : if(targetfilename_.length()>0) {
1302 2 : IFile gridfile; gridfile.open(targetfilename_);
1303 2 : std::string funcl=getLabel() + ".target";
1304 4 : TargetGrid_=GridBase::create(funcl,getArguments(),gridfile,false,false,true);
1305 2 : if(TargetGrid_->getDimension()!=getNumberOfArguments()) error("mismatch between dimensionality of input grid and number of arguments");
1306 4 : for(unsigned i=0; i<getNumberOfArguments(); ++i) {
1307 4 : if( getPntrToArgument(i)->isPeriodic()!=TargetGrid_->getIsPeriodic()[i] ) error("periodicity mismatch between arguments and input bias");
1308 : }
1309 2 : }
1310 :
1311 148 : if(getRestart()) {
1312 : // if this is a restart the neighbor list should be immediately updated
1313 37 : if(nlist_) nlist_update_=true;
1314 : // Calculate the Tiwary-Parrinello reweighting factor if we are restarting from previous hills
1315 37 : if(calc_rct_) computeReweightingFactor();
1316 : // Calculate all special bias quantities desired if restarting with nonzero bias.
1317 37 : if(calc_max_bias_) {
1318 0 : max_bias_ = BiasGrid_->getMaxValue();
1319 0 : getPntrToComponent("maxbias")->set(max_bias_);
1320 : }
1321 37 : if(calc_transition_bias_) {
1322 13 : transition_bias_ = getTransitionBarrierBias();
1323 26 : getPntrToComponent("transbias")->set(transition_bias_);
1324 : }
1325 : }
1326 :
1327 : // open grid file for writing
1328 148 : if(wgridstride_>0) {
1329 19 : gridfile_.link(*this);
1330 19 : if(walkers_mpi_) {
1331 0 : int r=0;
1332 0 : if(comm.Get_rank()==0) r=multi_sim_comm.Get_rank();
1333 0 : comm.Bcast(r,0);
1334 0 : if(r>0) gridfilename_="/dev/null";
1335 0 : gridfile_.enforceSuffix("");
1336 : }
1337 19 : if(mw_n_>1) gridfile_.enforceSuffix("");
1338 19 : gridfile_.open(gridfilename_);
1339 : }
1340 :
1341 : // open hills file for writing
1342 148 : hillsOfile_.link(*this);
1343 148 : if(walkers_mpi_) {
1344 38 : int r=0;
1345 38 : if(comm.Get_rank()==0) r=multi_sim_comm.Get_rank();
1346 38 : comm.Bcast(r,0);
1347 38 : if(r>0) ifilesnames_[mw_id_]="/dev/null";
1348 76 : hillsOfile_.enforceSuffix("");
1349 : }
1350 154 : if(mw_n_>1) hillsOfile_.enforceSuffix("");
1351 148 : hillsOfile_.open(ifilesnames_[mw_id_]);
1352 148 : if(fmt_.length()>0) hillsOfile_.fmtField(fmt_);
1353 148 : hillsOfile_.addConstantField("multivariate");
1354 148 : hillsOfile_.addConstantField("kerneltype");
1355 148 : if(doInt_) {
1356 4 : hillsOfile_.addConstantField("lower_int").printField("lower_int",lowI_);
1357 4 : hillsOfile_.addConstantField("upper_int").printField("upper_int",uppI_);
1358 : }
1359 : hillsOfile_.setHeavyFlush();
1360 : // output periodicities of variables
1361 415 : for(unsigned i=0; i<getNumberOfArguments(); ++i) hillsOfile_.setupPrintValue( getPntrToArgument(i) );
1362 :
1363 : bool concurrent=false;
1364 148 : const ActionSet&actionSet(plumed.getActionSet());
1365 1941 : for(const auto & p : actionSet) if(dynamic_cast<MetaD*>(p.get())) { concurrent=true; break; }
1366 148 : if(concurrent) log<<" You are using concurrent metadynamics\n";
1367 148 : if(rect_biasf_.size()>0) {
1368 18 : if(walkers_mpi_) {
1369 12 : log<<" You are using RECT in its 'altruistic' implementation\n";
1370 : }{
1371 18 : log<<" You are using RECT\n";
1372 : }
1373 : }
1374 :
1375 296 : log<<" Bibliography "<<plumed.cite("Laio and Parrinello, PNAS 99, 12562 (2002)");
1376 186 : if(welltemp_) log<<plumed.cite("Barducci, Bussi, and Parrinello, Phys. Rev. Lett. 100, 020603 (2008)");
1377 148 : if(tt_specs_.is_active) {
1378 6 : log << plumed.cite("Dama, Rotskoff, Parrinello, and Voth, J. Chem. Theory Comput. 10, 3626 (2014)");
1379 6 : log << plumed.cite("Dama, Parrinello, and Voth, Phys. Rev. Lett. 112, 240602 (2014)");
1380 : }
1381 192 : if(mw_n_>1||walkers_mpi_) log<<plumed.cite("Raiteri, Laio, Gervasio, Micheletti, and Parrinello, J. Phys. Chem. B 110, 3533 (2006)");
1382 169 : if(adaptive_!=FlexibleBin::none) log<<plumed.cite("Branduardi, Bussi, and Parrinello, J. Chem. Theory Comput. 8, 2247 (2012)");
1383 150 : if(doInt_) log<<plumed.cite("Baftizadeh, Cossio, Pietrucci, and Laio, Curr. Phys. Chem. 2, 79 (2012)");
1384 152 : if(acceleration_) log<<plumed.cite("Pratyush and Parrinello, Phys. Rev. Lett. 111, 230602 (2013)");
1385 154 : if(calc_rct_) log<<plumed.cite("Pratyush and Parrinello, J. Phys. Chem. B, 119, 736 (2015)");
1386 228 : if(concurrent || rect_biasf_.size()>0) log<<plumed.cite("Gil-Ley and Bussi, J. Chem. Theory Comput. 11, 1077 (2015)");
1387 160 : if(rect_biasf_.size()>0 && walkers_mpi_) log<<plumed.cite("Hosek, Toulcova, Bortolato, and Spiwok, J. Phys. Chem. B 120, 2209 (2016)");
1388 148 : if(targetfilename_.length()>0) {
1389 4 : log<<plumed.cite("White, Dama, and Voth, J. Chem. Theory Comput. 11, 2451 (2015)");
1390 4 : log<<plumed.cite("Marinelli and Faraldo-Gómez, Biophys. J. 108, 2779 (2015)");
1391 4 : log<<plumed.cite("Gil-Ley, Bottaro, and Bussi, J. Chem. Theory Comput. 12, 2790 (2016)");
1392 : }
1393 150 : if(freq_adaptive_) log<<plumed.cite("Wang, Valsson, Tiwary, Parrinello, and Lindorff-Larsen, J. Chem. Phys. 149, 072309 (2018)");
1394 148 : log<<"\n";
1395 396 : }
1396 :
1397 151 : void MetaD::readTemperingSpecs(TemperingSpecs &t_specs)
1398 : {
1399 : // Set global tempering parameters.
1400 151 : parse(t_specs.name_stem + "BIASFACTOR", t_specs.biasf);
1401 151 : if (t_specs.biasf != -1.0) {
1402 3 : if (kbt_ == 0.0) {
1403 0 : error("Unless the MD engine passes the temperature to plumed, with tempered metad you must specify it using TEMP");
1404 : }
1405 3 : if (t_specs.biasf == 1.0) {
1406 0 : error("A bias factor of 1 corresponds to zero delta T and zero hill size, so it is not allowed.");
1407 : }
1408 3 : t_specs.is_active = true;
1409 3 : parse(t_specs.name_stem + "BIASTHRESHOLD", t_specs.threshold);
1410 3 : if (t_specs.threshold < 0.0) {
1411 0 : error(t_specs.name + " bias threshold is nonsensical");
1412 : }
1413 3 : parse(t_specs.name_stem + "ALPHA", t_specs.alpha);
1414 3 : if (t_specs.alpha <= 0.0 || t_specs.alpha > 1.0) {
1415 0 : error(t_specs.name + " decay shape parameter alpha is nonsensical");
1416 : }
1417 : }
1418 151 : }
1419 :
1420 3 : void MetaD::logTemperingSpecs(const TemperingSpecs &t_specs)
1421 : {
1422 3 : log.printf(" %s bias factor %f\n", t_specs.name.c_str(), t_specs.biasf);
1423 3 : log.printf(" KbT %f\n", kbt_);
1424 3 : if (t_specs.threshold != 0.0) log.printf(" %s bias threshold %f\n", t_specs.name.c_str(), t_specs.threshold);
1425 3 : if (t_specs.alpha != 1.0) log.printf(" %s decay shape parameter alpha %f\n", t_specs.name.c_str(), t_specs.alpha);
1426 3 : }
1427 :
1428 6036 : void MetaD::readGaussians(IFile *ifile)
1429 : {
1430 6036 : unsigned ncv=getNumberOfArguments();
1431 6036 : std::vector<double> center(ncv);
1432 6036 : std::vector<double> sigma(ncv);
1433 : double height;
1434 : int nhills=0;
1435 6036 : bool multivariate=false;
1436 :
1437 : std::vector<Value> tmpvalues;
1438 18121 : for(unsigned j=0; j<getNumberOfArguments(); ++j) tmpvalues.push_back( Value( this, getPntrToArgument(j)->getName(), false ) );
1439 :
1440 8227 : while(scanOneHill(ifile,tmpvalues,center,sigma,height,multivariate))
1441 : {
1442 2191 : nhills++;
1443 : // note that for gamma=1 we store directly -F
1444 2191 : if(welltemp_ && biasf_>1.0) height*=(biasf_-1.0)/biasf_;
1445 2191 : addGaussian(Gaussian(multivariate,height,center,sigma));
1446 : }
1447 6036 : log.printf(" %d Gaussians read\n",nhills);
1448 12072 : }
1449 :
1450 2922 : void MetaD::writeGaussian(const Gaussian& hill, OFile&file)
1451 : {
1452 2922 : unsigned ncv=getNumberOfArguments();
1453 2922 : file.printField("time",getTimeStep()*getStep());
1454 8194 : for(unsigned i=0; i<ncv; ++i) {
1455 5272 : file.printField(getPntrToArgument(i),hill.center[i]);
1456 : }
1457 5844 : hillsOfile_.printField("kerneltype","stretched-gaussian");
1458 2922 : if(hill.multivariate) {
1459 892 : hillsOfile_.printField("multivariate","true");
1460 : Matrix<double> mymatrix(ncv,ncv);
1461 : unsigned k=0;
1462 1047 : for(unsigned i=0; i<ncv; i++) {
1463 1357 : for(unsigned j=i; j<ncv; j++) {
1464 : // recompose the full inverse matrix
1465 756 : mymatrix(i,j)=mymatrix(j,i)=hill.sigma[k];
1466 756 : k++;
1467 : }
1468 : }
1469 : // invert it
1470 : Matrix<double> invmatrix(ncv,ncv);
1471 446 : Invert(mymatrix,invmatrix);
1472 : // enforce symmetry
1473 1047 : for(unsigned i=0; i<ncv; i++) {
1474 1357 : for(unsigned j=i; j<ncv; j++) {
1475 756 : invmatrix(i,j)=invmatrix(j,i);
1476 : }
1477 : }
1478 :
1479 : // do cholesky so to have a "sigma like" number
1480 : Matrix<double> lower(ncv,ncv);
1481 446 : cholesky(invmatrix,lower);
1482 : // loop in band form
1483 1047 : for(unsigned i=0; i<ncv; i++) {
1484 1357 : for(unsigned j=0; j<ncv-i; j++) {
1485 1512 : file.printField("sigma_"+getPntrToArgument(j+i)->getName()+"_"+getPntrToArgument(j)->getName(),lower(j+i,j));
1486 : }
1487 : }
1488 : } else {
1489 4952 : hillsOfile_.printField("multivariate","false");
1490 7147 : for(unsigned i=0; i<ncv; ++i)
1491 9342 : file.printField("sigma_"+getPntrToArgument(i)->getName(),hill.sigma[i]);
1492 : }
1493 2922 : double height=hill.height;
1494 : // note that for gamma=1 we store directly -F
1495 2922 : if(welltemp_ && biasf_>1.0) height*=biasf_/(biasf_-1.0);
1496 5844 : file.printField("height",height).printField("biasf",biasf_);
1497 4431 : if(mw_n_>1) file.printField("clock",int(std::time(0)));
1498 2922 : file.printField();
1499 2922 : }
1500 :
1501 5275 : void MetaD::addGaussian(const Gaussian& hill)
1502 : {
1503 5275 : if(grid_) {
1504 640 : size_t ncv=getNumberOfArguments();
1505 640 : std::vector<unsigned> nneighb=getGaussianSupport(hill);
1506 640 : std::vector<Grid::index_t> neighbors=BiasGrid_->getNeighbors(hill.center,nneighb);
1507 640 : std::vector<double> der(ncv);
1508 640 : std::vector<double> xx(ncv);
1509 640 : if(comm.Get_size()==1) {
1510 : // for performance reasons and thread safety
1511 544 : std::vector<double> dp(ncv);
1512 55324 : for(size_t i=0; i<neighbors.size(); ++i) {
1513 54780 : Grid::index_t ineigh=neighbors[i];
1514 158922 : for(unsigned j=0; j<ncv; ++j) der[j]=0.0;
1515 54780 : BiasGrid_->getPoint(ineigh,xx);
1516 54780 : double bias=evaluateGaussianAndDerivatives(xx,hill,der,dp);
1517 54780 : BiasGrid_->addValueAndDerivatives(ineigh,bias,der);
1518 : }
1519 : } else {
1520 96 : unsigned stride=comm.Get_size();
1521 96 : unsigned rank=comm.Get_rank();
1522 96 : std::vector<double> allder(ncv*neighbors.size(),0.0);
1523 96 : std::vector<double> n_der(ncv,0.0);
1524 96 : std::vector<double> allbias(neighbors.size(),0.0);
1525 : // for performance reasons and thread safety
1526 96 : std::vector<double> dp(ncv);
1527 27148 : for(unsigned i=rank; i<neighbors.size(); i+=stride) {
1528 27052 : Grid::index_t ineigh=neighbors[i];
1529 81156 : for(unsigned j=0; j<ncv; ++j) n_der[j]=0.0;
1530 27052 : BiasGrid_->getPoint(ineigh,xx);
1531 27052 : allbias[i]=evaluateGaussianAndDerivatives(xx,hill,n_der,dp);
1532 81156 : for(unsigned j=0; j<ncv; j++) allder[ncv*i+j]=n_der[j];
1533 : }
1534 96 : comm.Sum(allbias);
1535 96 : comm.Sum(allder);
1536 103200 : for(unsigned i=0; i<neighbors.size(); ++i) {
1537 103104 : Grid::index_t ineigh=neighbors[i];
1538 309312 : for(unsigned j=0; j<ncv; ++j) der[j]=allder[ncv*i+j];
1539 103104 : BiasGrid_->addValueAndDerivatives(ineigh,allbias[i],der);
1540 : }
1541 : }
1542 4635 : } else hills_.push_back(hill);
1543 5275 : }
1544 :
1545 640 : std::vector<unsigned> MetaD::getGaussianSupport(const Gaussian& hill)
1546 : {
1547 : std::vector<unsigned> nneigh;
1548 : std::vector<double> cutoff;
1549 640 : unsigned ncv=getNumberOfArguments();
1550 :
1551 : // traditional or flexible hill?
1552 640 : if(hill.multivariate) {
1553 : unsigned k=0;
1554 : Matrix<double> mymatrix(ncv,ncv);
1555 0 : for(unsigned i=0; i<ncv; i++) {
1556 0 : for(unsigned j=i; j<ncv; j++) {
1557 : // recompose the full inverse matrix
1558 0 : mymatrix(i,j)=mymatrix(j,i)=hill.sigma[k];
1559 0 : k++;
1560 : }
1561 : }
1562 : // Reinvert so to have the ellipses
1563 : Matrix<double> myinv(ncv,ncv);
1564 0 : Invert(mymatrix,myinv);
1565 : Matrix<double> myautovec(ncv,ncv);
1566 0 : std::vector<double> myautoval(ncv); //should I take this or their square root?
1567 0 : diagMat(myinv,myautoval,myautovec);
1568 : double maxautoval=0.;
1569 : unsigned ind_maxautoval; ind_maxautoval=ncv;
1570 0 : for(unsigned i=0; i<ncv; i++) {
1571 0 : if(myautoval[i]>maxautoval) {maxautoval=myautoval[i]; ind_maxautoval=i;}
1572 : }
1573 0 : for(unsigned i=0; i<ncv; i++) {
1574 0 : cutoff.push_back(std::sqrt(2.0*dp2cutoff)*std::abs(std::sqrt(maxautoval)*myautovec(i,ind_maxautoval)));
1575 : }
1576 : } else {
1577 1618 : for(unsigned i=0; i<ncv; ++i) {
1578 978 : cutoff.push_back(std::sqrt(2.0*dp2cutoff)*hill.sigma[i]);
1579 : }
1580 : }
1581 :
1582 640 : if(doInt_) {
1583 2 : if(hill.center[0]+cutoff[0] > uppI_ || hill.center[0]-cutoff[0] < lowI_) {
1584 : // in this case, we updated the entire grid to avoid problems
1585 2 : return BiasGrid_->getNbin();
1586 : } else {
1587 0 : nneigh.push_back( static_cast<unsigned>(ceil(cutoff[0]/BiasGrid_->getDx()[0])) );
1588 : return nneigh;
1589 : }
1590 : } else {
1591 1614 : for(unsigned i=0; i<ncv; i++) {
1592 976 : nneigh.push_back( static_cast<unsigned>(ceil(cutoff[i]/BiasGrid_->getDx()[i])) );
1593 : }
1594 : }
1595 :
1596 : return nneigh;
1597 : }
1598 :
1599 285 : double MetaD::getBias(const std::vector<double>& cv)
1600 : {
1601 285 : double bias=0.0;
1602 285 : if(grid_) bias = BiasGrid_->getValue(cv);
1603 : else {
1604 82 : unsigned nt=OpenMP::getNumThreads();
1605 82 : unsigned stride=comm.Get_size();
1606 82 : unsigned rank=comm.Get_rank();
1607 :
1608 82 : if(!nlist_) {
1609 82 : #pragma omp parallel num_threads(nt)
1610 : {
1611 : #pragma omp for reduction(+:bias) nowait
1612 : for(unsigned i=rank; i<hills_.size(); i+=stride) bias+=evaluateGaussian(cv,hills_[i]);
1613 : }
1614 : } else {
1615 0 : #pragma omp parallel num_threads(nt)
1616 : {
1617 : #pragma omp for reduction(+:bias) nowait
1618 : for(unsigned i=rank; i<nlist_hills_.size(); i+=stride) bias+=evaluateGaussian(cv,nlist_hills_[i]);
1619 : }
1620 : }
1621 82 : comm.Sum(bias);
1622 : }
1623 :
1624 285 : return bias;
1625 : }
1626 :
1627 8395 : double MetaD::getBiasAndDerivatives(const std::vector<double>& cv, std::vector<double>& der)
1628 : {
1629 8395 : unsigned ncv=getNumberOfArguments();
1630 8395 : double bias=0.0;
1631 8395 : if(grid_) {
1632 1506 : std::vector<double> vder(ncv);
1633 1506 : bias=BiasGrid_->getValueAndDerivatives(cv,vder);
1634 3498 : for(unsigned i=0; i<ncv; i++) der[i]=vder[i];
1635 : } else {
1636 6889 : unsigned nt=OpenMP::getNumThreads();
1637 6889 : unsigned stride=comm.Get_size();
1638 6889 : unsigned rank=comm.Get_rank();
1639 :
1640 6889 : if(!nlist_) {
1641 6884 : if(hills_.size()<2*nt*stride||nt==1) {
1642 : // for performance reasons and thread safety
1643 2587 : std::vector<double> dp(ncv);
1644 6704 : for(unsigned i=rank; i<hills_.size(); i+=stride) {
1645 4117 : bias+=evaluateGaussianAndDerivatives(cv,hills_[i],der,dp);
1646 : }
1647 : } else {
1648 4297 : #pragma omp parallel num_threads(nt)
1649 : {
1650 : std::vector<double> omp_deriv(ncv,0.);
1651 : // for performance reasons and thread safety
1652 : std::vector<double> dp(ncv);
1653 : #pragma omp for reduction(+:bias) nowait
1654 : for(unsigned i=rank; i<hills_.size(); i+=stride) {
1655 : bias+=evaluateGaussianAndDerivatives(cv,hills_[i],omp_deriv,dp);
1656 : }
1657 : #pragma omp critical
1658 : for(unsigned i=0; i<ncv; i++) der[i]+=omp_deriv[i];
1659 : }
1660 : }
1661 : } else {
1662 5 : if(hills_.size()<2*nt*stride||nt==1) {
1663 : // for performance reasons and thread safety
1664 0 : std::vector<double> dp(ncv);
1665 0 : for(unsigned i=rank; i<nlist_hills_.size(); i+=stride) {
1666 0 : bias+=evaluateGaussianAndDerivatives(cv,nlist_hills_[i],der,dp);
1667 : }
1668 : } else {
1669 5 : #pragma omp parallel num_threads(nt)
1670 : {
1671 : std::vector<double> omp_deriv(ncv,0.);
1672 : // for performance reasons and thread safety
1673 : std::vector<double> dp(ncv);
1674 : #pragma omp for reduction(+:bias) nowait
1675 : for(unsigned i=rank; i<nlist_hills_.size(); i+=stride) {
1676 : bias+=evaluateGaussianAndDerivatives(cv,nlist_hills_[i],omp_deriv,dp);
1677 : }
1678 : #pragma omp critical
1679 : for(unsigned i=0; i<ncv; i++) der[i]+=omp_deriv[i];
1680 : }
1681 : }
1682 : }
1683 6889 : comm.Sum(bias);
1684 6889 : comm.Sum(der);
1685 : }
1686 :
1687 8395 : return bias;
1688 : }
1689 :
1690 0 : double MetaD::getGaussianNormalization(const Gaussian& hill)
1691 : {
1692 : double norm=1;
1693 0 : unsigned ncv=hill.center.size();
1694 :
1695 0 : if(hill.multivariate) {
1696 : // recompose the full sigma from the upper diag cholesky
1697 : unsigned k=0;
1698 : Matrix<double> mymatrix(ncv,ncv);
1699 0 : for(unsigned i=0; i<ncv; i++) {
1700 0 : for(unsigned j=i; j<ncv; j++) {
1701 0 : mymatrix(i,j)=mymatrix(j,i)=hill.sigma[k]; // recompose the full inverse matrix
1702 0 : k++;
1703 : }
1704 0 : double ldet; logdet( mymatrix, ldet );
1705 0 : norm = std::exp( ldet ); // Not sure here if mymatrix is sigma or inverse
1706 : }
1707 : } else {
1708 0 : for(unsigned i=0; i<hill.sigma.size(); i++) norm*=hill.sigma[i];
1709 : }
1710 :
1711 0 : return norm*std::pow(2*pi,static_cast<double>(ncv)/2.0);
1712 : }
1713 :
1714 192 : double MetaD::evaluateGaussian(const std::vector<double>& cv, const Gaussian& hill)
1715 : {
1716 192 : unsigned ncv=cv.size();
1717 :
1718 : // I use a pointer here because cv is const (and should be const)
1719 : // but when using doInt it is easier to locally replace cv[0] with
1720 : // the upper/lower limit in case it is out of range
1721 : double tmpcv[1];
1722 : const double *pcv=NULL; // pointer to cv
1723 192 : if(ncv>0) pcv=&cv[0];
1724 192 : if(doInt_) {
1725 0 : plumed_assert(ncv==1);
1726 0 : tmpcv[0]=cv[0];
1727 0 : if(cv[0]<lowI_) tmpcv[0]=lowI_;
1728 0 : if(cv[0]>uppI_) tmpcv[0]=uppI_;
1729 : pcv=&(tmpcv[0]);
1730 : }
1731 :
1732 : double dp2=0.0;
1733 192 : if(hill.multivariate) {
1734 : unsigned k=0;
1735 : // recompose the full sigma from the upper diag cholesky
1736 : Matrix<double> mymatrix(ncv,ncv);
1737 0 : for(unsigned i=0; i<ncv; i++) {
1738 0 : for(unsigned j=i; j<ncv; j++) {
1739 0 : mymatrix(i,j)=mymatrix(j,i)=hill.sigma[k]; // recompose the full inverse matrix
1740 0 : k++;
1741 : }
1742 : }
1743 0 : for(unsigned i=0; i<ncv; i++) {
1744 0 : double dp_i=difference(i,hill.center[i],pcv[i]);
1745 0 : for(unsigned j=i; j<ncv; j++) {
1746 0 : if(i==j) {
1747 0 : dp2+=dp_i*dp_i*mymatrix(i,j)*0.5;
1748 : } else {
1749 0 : double dp_j=difference(j,hill.center[j],pcv[j]);
1750 0 : dp2+=dp_i*dp_j*mymatrix(i,j);
1751 : }
1752 : }
1753 : }
1754 : } else {
1755 576 : for(unsigned i=0; i<ncv; i++) {
1756 384 : double dp=difference(i,hill.center[i],pcv[i])*hill.invsigma[i];
1757 384 : dp2+=dp*dp;
1758 : }
1759 192 : dp2*=0.5;
1760 : }
1761 :
1762 : double bias=0.0;
1763 192 : if(dp2<dp2cutoff) bias=hill.height*(stretchA*std::exp(-dp2)+stretchB);
1764 :
1765 192 : return bias;
1766 : }
1767 :
1768 2408873 : double MetaD::evaluateGaussianAndDerivatives(const std::vector<double>& cv, const Gaussian& hill, std::vector<double>& der, std::vector<double>& dp_)
1769 : {
1770 2408873 : unsigned ncv=cv.size();
1771 :
1772 : // I use a pointer here because cv is const (and should be const)
1773 : // but when using doInt it is easier to locally replace cv[0] with
1774 : // the upper/lower limit in case it is out of range
1775 : const double *pcv=NULL; // pointer to cv
1776 : double tmpcv[1]; // tmp array with cv (to be used with doInt_)
1777 2408873 : if(ncv>0) pcv=&cv[0];
1778 2408873 : if(doInt_) {
1779 602 : plumed_assert(ncv==1);
1780 602 : tmpcv[0]=cv[0];
1781 602 : if(cv[0]<lowI_) tmpcv[0]=lowI_;
1782 602 : if(cv[0]>uppI_) tmpcv[0]=uppI_;
1783 : pcv=&(tmpcv[0]);
1784 : }
1785 :
1786 : bool int_der=false;
1787 2408873 : if(doInt_) {
1788 602 : if(cv[0]<lowI_ || cv[0]>uppI_) int_der=true;
1789 : }
1790 :
1791 : double dp2=0.0;
1792 : double bias=0.0;
1793 2408873 : if(hill.multivariate) {
1794 : unsigned k=0;
1795 : // recompose the full sigma from the upper diag cholesky
1796 : Matrix<double> mymatrix(ncv,ncv);
1797 161635 : for(unsigned i=0; i<ncv; i++) {
1798 162513 : for(unsigned j=i; j<ncv; j++) {
1799 81476 : mymatrix(i,j)=mymatrix(j,i)=hill.sigma[k]; // recompose the full inverse matrix
1800 81476 : k++;
1801 : }
1802 : }
1803 161635 : for(unsigned i=0; i<ncv; i++) {
1804 81037 : dp_[i]=difference(i,hill.center[i],pcv[i]);
1805 162513 : for(unsigned j=i; j<ncv; j++) {
1806 81476 : if(i==j) {
1807 81037 : dp2+=dp_[i]*dp_[i]*mymatrix(i,j)*0.5;
1808 : } else {
1809 439 : double dp_j=difference(j,hill.center[j],pcv[j]);
1810 439 : dp2+=dp_[i]*dp_j*mymatrix(i,j);
1811 : }
1812 : }
1813 : }
1814 80598 : if(dp2<dp2cutoff) {
1815 77683 : bias=hill.height*std::exp(-dp2);
1816 77683 : if(!int_der) {
1817 155673 : for(unsigned i=0; i<ncv; i++) {
1818 : double tmp=0.0;
1819 156594 : for(unsigned j=0; j<ncv; j++) tmp += dp_[j]*mymatrix(i,j)*bias;
1820 77990 : der[i]-=tmp*stretchA;
1821 : }
1822 : } else {
1823 0 : for(unsigned i=0; i<ncv; i++) der[i]=0.;
1824 : }
1825 77683 : bias=stretchA*bias+hill.height*stretchB;
1826 : }
1827 : } else {
1828 6975183 : for(unsigned i=0; i<ncv; i++) {
1829 4646908 : dp_[i]=difference(i,hill.center[i],pcv[i])*hill.invsigma[i];
1830 4646908 : dp2+=dp_[i]*dp_[i];
1831 : }
1832 2328275 : dp2*=0.5;
1833 2328275 : if(dp2<dp2cutoff) {
1834 1356525 : bias=hill.height*std::exp(-dp2);
1835 1356525 : if(!int_der) {
1836 4060796 : for(unsigned i=0; i<ncv; i++) der[i]-=bias*dp_[i]*hill.invsigma[i]*stretchA;
1837 : } else {
1838 478 : for(unsigned i=0; i<ncv; i++) der[i]=0.;
1839 : }
1840 1356525 : bias=stretchA*bias+hill.height*stretchB;
1841 : }
1842 : }
1843 :
1844 2408873 : return bias;
1845 : }
1846 :
1847 2736 : double MetaD::getHeight(const std::vector<double>& cv)
1848 : {
1849 2736 : double height=height0_;
1850 2736 : if(welltemp_) {
1851 275 : double vbias = getBias(cv);
1852 275 : if(biasf_>1.0) {
1853 259 : height = height0_*std::exp(-vbias/(kbt_*(biasf_-1.0)));
1854 : } else {
1855 : // notice that if gamma=1 we store directly -F
1856 16 : height = height0_*std::exp(-vbias/kbt_);
1857 : }
1858 : }
1859 2736 : if(dampfactor_>0.0) {
1860 18 : plumed_assert(BiasGrid_);
1861 18 : double m=BiasGrid_->getMaxValue();
1862 18 : height*=std::exp(-m/(kbt_*(dampfactor_)));
1863 : }
1864 2736 : if (tt_specs_.is_active) {
1865 60 : double vbarrier = transition_bias_;
1866 60 : temperHeight(height, tt_specs_, vbarrier);
1867 : }
1868 2736 : if(TargetGrid_) {
1869 18 : double f=TargetGrid_->getValue(cv)-TargetGrid_->getMaxValue();
1870 18 : height*=std::exp(f/kbt_);
1871 : }
1872 2736 : return height;
1873 : }
1874 :
1875 60 : void MetaD::temperHeight(double& height, const TemperingSpecs& t_specs, const double tempering_bias)
1876 : {
1877 60 : if (t_specs.alpha == 1.0) {
1878 80 : height *= std::exp(-std::max(0.0, tempering_bias - t_specs.threshold) / (kbt_ * (t_specs.biasf - 1.0)));
1879 : } else {
1880 40 : height *= std::pow(1 + (1 - t_specs.alpha) / t_specs.alpha * std::max(0.0, tempering_bias - t_specs.threshold) / (kbt_ * (t_specs.biasf - 1.0)), - t_specs.alpha / (1 - t_specs.alpha));
1881 : }
1882 60 : }
1883 :
1884 8435 : void MetaD::calculate()
1885 : {
1886 : // this is because presently there is no way to properly pass information
1887 : // on adaptive hills (diff) after exchanges:
1888 8435 : if(adaptive_==FlexibleBin::diffusion && getExchangeStep()) error("ADAPTIVE=DIFF is not compatible with replica exchange");
1889 :
1890 8435 : const unsigned ncv=getNumberOfArguments();
1891 8435 : std::vector<double> cv(ncv);
1892 21082 : for(unsigned i=0; i<ncv; ++i) cv[i]=getArgument(i);
1893 :
1894 8435 : if(nlist_) {
1895 5 : nlist_steps_++;
1896 5 : if(getExchangeStep()) nlist_update_=true;
1897 : else {
1898 11 : for(unsigned i=0; i<ncv; ++i) {
1899 8 : double d = difference(i, cv[i], nlist_center_[i]);
1900 8 : double nk_dist2 = d*d/nlist_dev2_[i];
1901 8 : if(nk_dist2>nlist_param_[1]) {nlist_update_=true; break;}
1902 : }
1903 : }
1904 5 : if(nlist_update_) updateNlist();
1905 : }
1906 :
1907 : double ene = 0.;
1908 8435 : std::vector<double> der(ncv,0.);
1909 8435 : if(biasf_!=1.0) ene = getBiasAndDerivatives(cv,der);
1910 8435 : setBias(ene);
1911 21082 : for(unsigned i=0; i<ncv; i++) setOutputForce(i,-der[i]);
1912 :
1913 8440 : if(calc_work_) getPntrToComponent("work")->set(work_);
1914 8545 : if(calc_rct_) getPntrToComponent("rbias")->set(ene - reweight_factor_);
1915 : // calculate the acceleration factor
1916 8435 : if(acceleration_&&!isFirstStep_) {
1917 329 : acc_ += static_cast<double>(getStride()) * std::exp(ene/(kbt_));
1918 329 : const double mean_acc = acc_/((double) getStep());
1919 329 : getPntrToComponent("acc")->set(mean_acc);
1920 8435 : } else if (acceleration_ && isFirstStep_ && acc_restart_mean_ > 0.0) {
1921 2 : acc_ = acc_restart_mean_ * static_cast<double>(getStep());
1922 2 : if(freq_adaptive_) {
1923 : // has to be done here if restarting, as the acc is not defined before
1924 1 : updateFrequencyAdaptiveStride();
1925 : }
1926 : }
1927 8435 : }
1928 :
1929 6239 : void MetaD::update()
1930 : {
1931 : // adding hills criteria (could be more complex though)
1932 : bool nowAddAHill;
1933 6239 : if(getStep()%current_stride_==0 && !isFirstStep_) nowAddAHill=true;
1934 : else {
1935 : nowAddAHill=false;
1936 3503 : isFirstStep_=false;
1937 : }
1938 :
1939 6239 : unsigned ncv=getNumberOfArguments();
1940 6239 : std::vector<double> cv(ncv);
1941 16690 : for(unsigned i=0; i<ncv; ++i) cv[i] = getArgument(i);
1942 :
1943 : double vbias=0.;
1944 6239 : if(calc_work_) vbias=getBias(cv);
1945 :
1946 : // if you use adaptive, call the FlexibleBin
1947 : bool multivariate=false;
1948 6239 : if(adaptive_!=FlexibleBin::none) {
1949 778 : flexbin_->update(nowAddAHill);
1950 : multivariate=true;
1951 : }
1952 :
1953 : std::vector<double> thissigma;
1954 6239 : if(nowAddAHill) {
1955 : // add a Gaussian
1956 2736 : double height=getHeight(cv);
1957 : // returns upper diagonal inverse
1958 3110 : if(adaptive_!=FlexibleBin::none) thissigma=flexbin_->getInverseMatrix();
1959 : // returns normal sigma
1960 2362 : else thissigma=sigma0_;
1961 :
1962 : // In case we use walkers_mpi, it is now necessary to communicate with other replicas.
1963 2736 : if(walkers_mpi_) {
1964 : // Allocate arrays to store all walkers hills
1965 174 : std::vector<double> all_cv(mpi_nw_*ncv,0.0);
1966 174 : std::vector<double> all_sigma(mpi_nw_*thissigma.size(),0.0);
1967 174 : std::vector<double> all_height(mpi_nw_,0.0);
1968 174 : std::vector<int> all_multivariate(mpi_nw_,0);
1969 174 : if(comm.Get_rank()==0) {
1970 : // Communicate (only root)
1971 99 : multi_sim_comm.Allgather(cv,all_cv);
1972 99 : multi_sim_comm.Allgather(thissigma,all_sigma);
1973 : // notice that if gamma=1 we store directly -F so this scaling is not necessary:
1974 99 : multi_sim_comm.Allgather(height*(biasf_>1.0?biasf_/(biasf_-1.0):1.0),all_height);
1975 99 : multi_sim_comm.Allgather(int(multivariate),all_multivariate);
1976 : }
1977 : // Share info with group members
1978 174 : comm.Bcast(all_cv,0);
1979 174 : comm.Bcast(all_sigma,0);
1980 174 : comm.Bcast(all_height,0);
1981 174 : comm.Bcast(all_multivariate,0);
1982 :
1983 : // Flying Gaussian
1984 174 : if (flying_) {
1985 54 : hills_.clear();
1986 54 : comm.Barrier();
1987 : }
1988 :
1989 696 : for(unsigned i=0; i<mpi_nw_; i++) {
1990 : // actually add hills one by one
1991 522 : std::vector<double> cv_now(ncv);
1992 522 : std::vector<double> sigma_now(thissigma.size());
1993 1566 : for(unsigned j=0; j<ncv; j++) cv_now[j]=all_cv[i*ncv+j];
1994 1674 : for(unsigned j=0; j<thissigma.size(); j++) sigma_now[j]=all_sigma[i*thissigma.size()+j];
1995 : // notice that if gamma=1 we store directly -F so this scaling is not necessary:
1996 522 : double fact=(biasf_>1.0?(biasf_-1.0)/biasf_:1.0);
1997 522 : Gaussian newhill=Gaussian(all_multivariate[i],all_height[i]*fact,cv_now,sigma_now);
1998 522 : addGaussian(newhill);
1999 522 : if(!flying_) writeGaussian(newhill,hillsOfile_);
2000 522 : }
2001 : } else {
2002 2562 : Gaussian newhill=Gaussian(multivariate,height,cv,thissigma);
2003 2562 : addGaussian(newhill);
2004 2562 : writeGaussian(newhill,hillsOfile_);
2005 2562 : }
2006 :
2007 : // this is to update the hills neighbor list
2008 2736 : if(nlist_) nlist_update_=true;
2009 : }
2010 :
2011 : // this should be outside of the if block in case
2012 : // mw_rstride_ is not a multiple of stride_
2013 6239 : if(mw_n_>1 && getStep()%mw_rstride_==0) hillsOfile_.flush();
2014 :
2015 6239 : if(calc_work_) {
2016 5 : if(nlist_) updateNlist();
2017 5 : double vbias1=getBias(cv);
2018 5 : work_+=vbias1-vbias;
2019 : }
2020 :
2021 : // dump grid on file
2022 6239 : if(wgridstride_>0&&(getStep()%wgridstride_==0||getCPT())) {
2023 : // in case old grids are stored, a sequence of grids should appear
2024 : // this call results in a repetition of the header:
2025 91 : if(storeOldGrids_) gridfile_.clearFields();
2026 : // in case only latest grid is stored, file should be rewound
2027 : // this will overwrite previously written grids
2028 : else {
2029 51 : int r = 0;
2030 51 : if(walkers_mpi_) {
2031 0 : if(comm.Get_rank()==0) r=multi_sim_comm.Get_rank();
2032 0 : comm.Bcast(r,0);
2033 : }
2034 51 : if(r==0) gridfile_.rewind();
2035 : }
2036 91 : BiasGrid_->writeToFile(gridfile_);
2037 : // if a single grid is stored, it is necessary to flush it, otherwise
2038 : // the file might stay empty forever (when a single grid is not large enough to
2039 : // trigger flushing from the operating system).
2040 : // on the other hand, if grids are stored one after the other this is
2041 : // no necessary, and we leave the flushing control to the user as usual
2042 : // (with FLUSH keyword)
2043 91 : if(!storeOldGrids_) gridfile_.flush();
2044 : }
2045 :
2046 : // if multiple walkers and time to read Gaussians
2047 6239 : if(mw_n_>1 && getStep()%mw_rstride_==0) {
2048 12048 : for(int i=0; i<mw_n_; ++i) {
2049 : // don't read your own Gaussians
2050 9036 : if(i==mw_id_) continue;
2051 : // if the file is not open yet
2052 6024 : if(!(ifiles_[i]->isOpen())) {
2053 : // check if it exists now and open it!
2054 7 : if(ifiles_[i]->FileExist(ifilesnames_[i])) {
2055 7 : ifiles_[i]->open(ifilesnames_[i]);
2056 7 : ifiles_[i]->reset(false);
2057 : }
2058 : // otherwise read the new Gaussians
2059 : } else {
2060 6017 : log.printf(" Reading hills from %s:",ifilesnames_[i].c_str());
2061 6017 : readGaussians(ifiles_[i].get());
2062 6017 : ifiles_[i]->reset(false);
2063 : }
2064 : }
2065 : // this is to update the hills neighbor list
2066 3012 : if(nlist_) nlist_update_=true;
2067 : }
2068 :
2069 : // Recalculate special bias quantities whenever the bias has been changed by the update.
2070 6239 : bool bias_has_changed = (nowAddAHill || (mw_n_ > 1 && getStep() % mw_rstride_ == 0));
2071 6239 : if (calc_rct_ && bias_has_changed && getStep()%(stride_*rct_ustride_)==0) computeReweightingFactor();
2072 6239 : if (calc_max_bias_ && bias_has_changed) {
2073 0 : max_bias_ = BiasGrid_->getMaxValue();
2074 0 : getPntrToComponent("maxbias")->set(max_bias_);
2075 : }
2076 6239 : if (calc_transition_bias_ && bias_has_changed) {
2077 260 : transition_bias_ = getTransitionBarrierBias();
2078 520 : getPntrToComponent("transbias")->set(transition_bias_);
2079 : }
2080 :
2081 : // Frequency adaptive metadynamics - update hill addition frequency
2082 6239 : if(freq_adaptive_ && getStep()%fa_update_frequency_==0) {
2083 151 : updateFrequencyAdaptiveStride();
2084 : }
2085 6239 : }
2086 :
2087 : /// takes a pointer to the file and a template std::string with values v and gives back the next center, sigma and height
2088 8227 : bool MetaD::scanOneHill(IFile* ifile, std::vector<Value>& tmpvalues, std::vector<double>& center, std::vector<double>& sigma, double& height, bool& multivariate)
2089 : {
2090 : double dummy;
2091 8227 : multivariate=false;
2092 16454 : if(ifile->scanField("time",dummy)) {
2093 2191 : unsigned ncv=tmpvalues.size();
2094 6527 : for(unsigned i=0; i<ncv; ++i) {
2095 4336 : ifile->scanField( &tmpvalues[i] );
2096 4336 : if( tmpvalues[i].isPeriodic() && ! getPntrToArgument(i)->isPeriodic() ) {
2097 0 : error("in hills file periodicity for variable " + tmpvalues[i].getName() + " does not match periodicity in input");
2098 4336 : } else if( tmpvalues[i].isPeriodic() ) {
2099 0 : std::string imin, imax; tmpvalues[i].getDomain( imin, imax );
2100 0 : std::string rmin, rmax; getPntrToArgument(i)->getDomain( rmin, rmax );
2101 0 : if( imin!=rmin || imax!=rmax ) {
2102 0 : error("in hills file periodicity for variable " + tmpvalues[i].getName() + " does not match periodicity in input");
2103 : }
2104 : }
2105 4336 : center[i]=tmpvalues[i].get();
2106 : }
2107 : // scan for kerneltype
2108 2191 : std::string ktype="stretched-gaussian";
2109 6564 : if( ifile->FieldExist("kerneltype") ) ifile->scanField("kerneltype",ktype);
2110 2191 : if( ktype=="gaussian" ) {
2111 12 : noStretchWarning();
2112 2179 : } else if( ktype!="stretched-gaussian") {
2113 0 : error("non Gaussian kernels are not supported in MetaD");
2114 : }
2115 : // scan for multivariate label: record the actual file position so to eventually rewind
2116 : std::string sss;
2117 4382 : ifile->scanField("multivariate",sss);
2118 2191 : if(sss=="true") multivariate=true;
2119 2191 : else if(sss=="false") multivariate=false;
2120 0 : else plumed_merror("cannot parse multivariate = "+ sss);
2121 2191 : if(multivariate) {
2122 0 : sigma.resize(ncv*(ncv+1)/2);
2123 : Matrix<double> upper(ncv,ncv);
2124 : Matrix<double> lower(ncv,ncv);
2125 0 : for(unsigned i=0; i<ncv; i++) {
2126 0 : for(unsigned j=0; j<ncv-i; j++) {
2127 0 : ifile->scanField("sigma_"+getPntrToArgument(j+i)->getName()+"_"+getPntrToArgument(j)->getName(),lower(j+i,j));
2128 0 : upper(j,j+i)=lower(j+i,j);
2129 : }
2130 : }
2131 : Matrix<double> mymult(ncv,ncv);
2132 : Matrix<double> invmatrix(ncv,ncv);
2133 0 : mult(lower,upper,mymult);
2134 : // now invert and get the sigmas
2135 0 : Invert(mymult,invmatrix);
2136 : // put the sigmas in the usual order: upper diagonal (this time in normal form and not in band form)
2137 : unsigned k=0;
2138 0 : for(unsigned i=0; i<ncv; i++) {
2139 0 : for(unsigned j=i; j<ncv; j++) {
2140 0 : sigma[k]=invmatrix(i,j);
2141 0 : k++;
2142 : }
2143 : }
2144 : } else {
2145 6527 : for(unsigned i=0; i<ncv; ++i) {
2146 8672 : ifile->scanField("sigma_"+getPntrToArgument(i)->getName(),sigma[i]);
2147 : }
2148 : }
2149 :
2150 2191 : ifile->scanField("height",height);
2151 2191 : ifile->scanField("biasf",dummy);
2152 6390 : if(ifile->FieldExist("clock")) ifile->scanField("clock",dummy);
2153 4382 : if(ifile->FieldExist("lower_int")) ifile->scanField("lower_int",dummy);
2154 4382 : if(ifile->FieldExist("upper_int")) ifile->scanField("upper_int",dummy);
2155 2191 : ifile->scanField();
2156 : return true;
2157 : } else {
2158 : return false;
2159 : }
2160 : }
2161 :
2162 102 : void MetaD::computeReweightingFactor()
2163 : {
2164 102 : if(biasf_==1.0) { // in this case we have no bias, so reweight factor is 0.0
2165 0 : getPntrToComponent("rct")->set(0.0);
2166 0 : return;
2167 : }
2168 :
2169 102 : double Z_0=0; //proportional to the integral of exp(-beta*F)
2170 102 : double Z_V=0; //proportional to the integral of exp(-beta*(F+V))
2171 102 : double minusBetaF=biasf_/(biasf_-1.)/kbt_;
2172 102 : double minusBetaFplusV=1./(biasf_-1.)/kbt_;
2173 102 : if (biasf_==-1.0) { //non well-tempered case
2174 0 : minusBetaF=1./kbt_;
2175 : minusBetaFplusV=0;
2176 : }
2177 102 : max_bias_=BiasGrid_->getMaxValue(); //to avoid exp overflow
2178 :
2179 102 : const unsigned rank=comm.Get_rank();
2180 102 : const unsigned stride=comm.Get_size();
2181 920504 : for (Grid::index_t t=rank; t<BiasGrid_->getSize(); t+=stride) {
2182 920402 : const double val=BiasGrid_->getValue(t);
2183 920402 : Z_0+=std::exp(minusBetaF*(val-max_bias_));
2184 920402 : Z_V+=std::exp(minusBetaFplusV*(val-max_bias_));
2185 : }
2186 102 : comm.Sum(Z_0);
2187 102 : comm.Sum(Z_V);
2188 :
2189 102 : reweight_factor_=kbt_*std::log(Z_0/Z_V)+max_bias_;
2190 204 : getPntrToComponent("rct")->set(reweight_factor_);
2191 : }
2192 :
2193 273 : double MetaD::getTransitionBarrierBias()
2194 : {
2195 : // If there is only one well of interest, return the bias at that well point.
2196 273 : if (transitionwells_.size() == 1) {
2197 0 : double tb_bias = getBias(transitionwells_[0]);
2198 0 : return tb_bias;
2199 :
2200 : // Otherwise, check for the least barrier bias between all pairs of wells.
2201 : // Note that because the paths can be considered edges between the wells' nodes
2202 : // to make a graph and the path barriers satisfy certain cycle inequalities, it
2203 : // is sufficient to look at paths corresponding to a minimal spanning tree of the
2204 : // overall graph rather than examining every edge in the graph.
2205 : // For simplicity, I chose the star graph with center well 0 as the spanning tree.
2206 : // It is most efficient to start the path searches from the wells that are
2207 : // expected to be sampled last, so transitionwell_[0] should correspond to the
2208 : // starting well. With this choice the searches will terminate in one step until
2209 : // transitionwell_[1] is sampled.
2210 : } else {
2211 : double least_transition_bias;
2212 273 : std::vector<double> sink = transitionwells_[0];
2213 273 : std::vector<double> source = transitionwells_[1];
2214 273 : least_transition_bias = BiasGrid_->findMaximalPathMinimum(source, sink);
2215 273 : for (unsigned i = 2; i < transitionwells_.size(); i++) {
2216 0 : if (least_transition_bias == 0.0) {
2217 : break;
2218 : }
2219 0 : source = transitionwells_[i];
2220 0 : double curr_transition_bias = BiasGrid_->findMaximalPathMinimum(source, sink);
2221 0 : least_transition_bias = fmin(curr_transition_bias, least_transition_bias);
2222 : }
2223 : return least_transition_bias;
2224 : }
2225 : }
2226 :
2227 154 : void MetaD::updateFrequencyAdaptiveStride()
2228 : {
2229 154 : plumed_massert(freq_adaptive_,"should only be used if frequency adaptive metadynamics is enabled");
2230 154 : plumed_massert(acceleration_,"frequency adaptive metadynamics can only be used if the acceleration factor is calculated");
2231 154 : const double mean_acc = acc_/((double) getStep());
2232 154 : int tmp_stride= stride_*floor((mean_acc/fa_min_acceleration_)+0.5);
2233 154 : if(mean_acc >= fa_min_acceleration_) {
2234 129 : if(tmp_stride > current_stride_) {current_stride_ = tmp_stride;}
2235 : }
2236 154 : if(fa_max_stride_!=0 && current_stride_>fa_max_stride_) {
2237 0 : current_stride_=fa_max_stride_;
2238 : }
2239 154 : getPntrToComponent("pace")->set(current_stride_);
2240 154 : }
2241 :
2242 8435 : bool MetaD::checkNeedsGradients()const
2243 : {
2244 8435 : if(adaptive_==FlexibleBin::geometry) {
2245 192 : if(getStep()%stride_==0 && !isFirstStep_) return true;
2246 109 : else return false;
2247 : } else return false;
2248 : }
2249 :
2250 4 : void MetaD::updateNlist()
2251 : {
2252 : // no need to check for neighbors
2253 4 : if(hills_.size()==0) return;
2254 :
2255 : // here we generate the neighbor list
2256 4 : nlist_hills_.clear();
2257 : std::vector<Gaussian> local_flat_nl;
2258 4 : unsigned nt=OpenMP::getNumThreads();
2259 4 : if(hills_.size()<2*nt) nt=1;
2260 4 : #pragma omp parallel num_threads(nt)
2261 : {
2262 : std::vector<Gaussian> private_flat_nl;
2263 : #pragma omp for nowait
2264 : for(unsigned k=0; k<hills_.size(); k++)
2265 : {
2266 : double dist2=0;
2267 : for(unsigned i=0; i<getNumberOfArguments(); i++)
2268 : {
2269 : const double d=difference(i,getArgument(i),hills_[k].center[i])/hills_[k].sigma[i];
2270 : dist2+=d*d;
2271 : }
2272 : if(dist2<=nlist_param_[0]*dp2cutoff) private_flat_nl.push_back(hills_[k]);
2273 : }
2274 : #pragma omp critical
2275 : local_flat_nl.insert(local_flat_nl.end(), private_flat_nl.begin(), private_flat_nl.end());
2276 : }
2277 4 : nlist_hills_ = local_flat_nl;
2278 :
2279 : // here we set some properties that are used to decide when to update it again
2280 12 : for(unsigned i=0; i<getNumberOfArguments(); i++) nlist_center_[i]=getArgument(i);
2281 : std::vector<double> dev2;
2282 4 : dev2.resize(getNumberOfArguments(),0);
2283 46 : for(unsigned k=0; k<nlist_hills_.size(); k++)
2284 : {
2285 126 : for(unsigned i=0; i<getNumberOfArguments(); i++)
2286 : {
2287 84 : const double d=difference(i,getArgument(i),nlist_hills_[k].center[i]);
2288 84 : dev2[i]+=d*d;
2289 : }
2290 : }
2291 12 : for(unsigned i=0; i<getNumberOfArguments(); i++) {
2292 8 : if(dev2[i]>0.) nlist_dev2_[i]=dev2[i]/static_cast<double>(nlist_hills_.size());
2293 0 : else nlist_dev2_[i]=hills_.back().sigma[i]*hills_.back().sigma[i];
2294 : }
2295 :
2296 : // we are done
2297 4 : getPntrToComponent("nlker")->set(nlist_hills_.size());
2298 4 : getPntrToComponent("nlsteps")->set(nlist_steps_);
2299 4 : nlist_steps_=0;
2300 4 : nlist_update_=false;
2301 4 : }
2302 :
2303 : }
2304 : }
|
