Line data Source code
1 : /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 : Copyright (c) 2016-2021 The VES code team
3 : (see the PEOPLE-VES file at the root of this folder for a list of names)
4 :
5 : See http://www.ves-code.org for more information.
6 :
7 : This file is part of VES code module.
8 :
9 : The VES code module 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 : The VES code module 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 the VES code module. If not, see <http://www.gnu.org/licenses/>.
21 : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
22 :
23 : #include "bias/Bias.h"
24 : #include "core/PlumedMain.h"
25 : #include "core/ActionRegister.h"
26 : #include "tools/Communicator.h"
27 : #include "tools/Grid.h"
28 : #include "tools/File.h"
29 : //#include <algorithm> //std::fill
30 :
31 : namespace PLMD {
32 : namespace ves {
33 :
34 : //+PLUMEDOC VES_BIAS VES_DELTA_F
35 : /*
36 : Implementation of VES Delta F method
37 :
38 : Implementation of VES\f$\Delta F\f$ method \cite Invernizzi2019vesdeltaf (step two only).
39 :
40 : \warning
41 : Notice that this is a stand-alone bias Action, it does not need any of the other VES module components
42 :
43 : First you should create some estimate of the local free energy basins of your system,
44 : using e.g. multiple \ref METAD short runs, and combining them with the \ref sum_hills utility.
45 : Once you have them, you can use this bias Action to perform the VES optimization part of the method.
46 :
47 : These \f$N+1\f$ local basins are used to model the global free energy.
48 : In particular, given the conditional probabilities \f$P(\mathbf{s}|i)\propto e^{-\beta F_i(\mathbf{s})}\f$
49 : and the probabilities of being in a given basin \f$P_i\f$, we can write:
50 : \f[
51 : e^{-\beta F(\mathbf{s})}\propto P(\mathbf{s})=\sum_{i=0}^N P(\mathbf{s}|i)P_i \, .
52 : \f]
53 : We use this free energy model and the chosen bias factor \f$\gamma\f$ to build the bias potential:
54 : \f$V(\mathbf{s})=-(1-1/\gamma)F(\mathbf{s})\f$.
55 : Or, more explicitly:
56 : \f[
57 : V(\mathbf{s})=(1-1/\gamma)\frac{1}{\beta}\log\left[e^{-\beta F_0(\mathbf{s})}
58 : +\sum_{i=1}^{N} e^{-\beta F_i(\mathbf{s})} e^{-\beta \alpha_i}\right] \, ,
59 : \f]
60 : where the parameters \f$\boldsymbol{\alpha}\f$ are the \f$N\f$ free energy differences (see below) from the \f$F_0\f$ basin.
61 :
62 : By default the \f$F_i(\mathbf{s})\f$ are shifted so that \f$\min[F_i(\mathbf{s})]=0\f$ for all \f$i=\{0,...,N\}\f$.
63 : In this case the optimization parameters \f$\alpha_i\f$ are the difference in height between the minima of the basins.
64 : Using the keyword `NORMALIZE`, you can also decide to normalize the local free energies so that
65 : \f$\int d\mathbf{s}\, e^{-\beta F_i(\mathbf{s})}=1\f$.
66 : In this case the parameters will represent not the difference in height (which depends on the chosen CVs),
67 : but the actual free energy difference, \f$\alpha_i=\Delta F_i\f$.
68 :
69 : However, as discussed in Ref. \cite Invernizzi2019vesdeltaf, a better estimate of \f$\Delta F_i\f$ should be obtained through the reweighting procedure.
70 :
71 : \par Examples
72 :
73 : The following performs the optimization of the free energy difference between two metastable basins:
74 :
75 : \plumedfile
76 : cv: DISTANCE ATOMS=1,2
77 : ves: VES_DELTA_F ...
78 : ARG=cv
79 : TEMP=300
80 : FILE_F0=fesA.data
81 : FILE_F1=fesB.data
82 : BIASFACTOR=10.0
83 : M_STEP=0.1
84 : AV_STRIDE=500
85 : PRINT_STRIDE=100
86 : ...
87 : PRINT FMT=%g STRIDE=500 FILE=Colvar.data ARG=cv,ves.bias,ves.rct
88 : \endplumedfile
89 :
90 : The local FES files can be obtained as described in Sec. 4.2 of Ref. \cite Invernizzi2019vesdeltaf, i.e. for example:
91 : - run 4 independent metad runs, all starting from basin A, and kill them as soon as they make the first transition (see e.g. \ref COMMITTOR)
92 : - \verbatim cat HILLS* > all_HILLS \endverbatim
93 : - \verbatim plumed sum_hills --hills all_HILLS --outfile all_fesA.dat --mintozero --min 0 --max 1 --bin 100 \endverbatim
94 : - \verbatim awk -v n_rep=4 '{if($1!="#!" && $1!="") {for(i=1+(NF-1)/2; i<=NF; i++) $i/=n_rep;} print $0}' all_fesA.dat > fesA.data \endverbatim
95 :
96 : The header of both FES files must be identical, and should be similar to the following:
97 :
98 : \auxfile{fesA.data}
99 : #! FIELDS cv file.free der_cv
100 : #! SET min_cv 0
101 : #! SET max_cv 1
102 : #! SET nbins_cv 100
103 : #! SET periodic_cv false
104 : 0 0 0
105 : \endauxfile
106 : \auxfile{fesB.data}
107 : #! FIELDS cv file.free der_cv
108 : #! SET min_cv 0
109 : #! SET max_cv 1
110 : #! SET nbins_cv 100
111 : #! SET periodic_cv false
112 : 0 0 0
113 : \endauxfile
114 :
115 : */
116 : //+ENDPLUMEDOC
117 :
118 : class VesDeltaF : public bias::Bias {
119 :
120 : private:
121 : double beta_;
122 : unsigned NumParallel_;
123 : unsigned rank_;
124 : unsigned NumWalkers_;
125 : bool isFirstStep_;
126 : bool afterCalculate_;
127 :
128 : //prob
129 : double tot_prob_;
130 : std::vector<double> prob_;
131 : std::vector< std::vector<double> > der_prob_;
132 :
133 : //local basins
134 : std::vector< std::unique_ptr<Grid> > grid_p_; //pointers because of GridBase::create
135 : std::vector<double> norm_;
136 :
137 : //optimizer-related stuff
138 : long long unsigned mean_counter_;
139 : unsigned mean_weight_tau_;
140 : unsigned alpha_size_;
141 : unsigned sym_alpha_size_;
142 : std::vector<double> mean_alpha_;
143 : std::vector<double> inst_alpha_;
144 : std::vector<double> past_increment2_;
145 : double minimization_step_;
146 : bool damping_off_;
147 : //'tg' -> 'target distribution'
148 : double inv_gamma_;
149 : unsigned tg_counter_;
150 : unsigned tg_stride_;
151 : std::vector<double> tg_dV_dAlpha_;
152 : std::vector<double> tg_d2V_dAlpha2_;
153 : //'av' -> 'ensemble average'
154 : unsigned av_counter_;
155 : unsigned av_stride_;
156 : std::vector<double> av_dV_dAlpha_;
157 : std::vector<double> av_dV_dAlpha_prod_;
158 : std::vector<double> av_d2V_dAlpha2_;
159 : //printing
160 : unsigned print_stride_;
161 : OFile alphaOfile_;
162 : //other
163 : std::vector<double> exp_alpha_;
164 : std::vector<double> prev_exp_alpha_;
165 : double work_;
166 :
167 : //functions
168 : void update_alpha();
169 : void update_tg_and_rct();
170 : inline unsigned get_index(const unsigned, const unsigned) const;
171 :
172 : public:
173 : explicit VesDeltaF(const ActionOptions&);
174 : void calculate() override;
175 : void update() override;
176 : static void registerKeywords(Keywords& keys);
177 : };
178 :
179 : PLUMED_REGISTER_ACTION(VesDeltaF,"VES_DELTA_F")
180 :
181 6 : void VesDeltaF::registerKeywords(Keywords& keys) {
182 6 : Bias::registerKeywords(keys);
183 6 : keys.add("optional","TEMP","temperature is compulsory, but it can be sometimes fetched from the MD engine");
184 : //local free energies
185 6 : keys.add("numbered","FILE_F","names of files containing local free energies and derivatives. "
186 : "The first one, FILE_F0, is used as reference for all the free energy differences.");
187 12 : keys.reset_style("FILE_F","compulsory");
188 6 : keys.addFlag("NORMALIZE",false,"normalize all local free energies so that alpha will be (approx) Delta F");
189 6 : keys.addFlag("NO_MINTOZERO",false,"leave local free energies as provided, without shifting them to zero min");
190 : //target distribution
191 6 : keys.add("compulsory","BIASFACTOR","0","the gamma bias factor used for well-tempered target p(s)."
192 : " Set to 0 for non-tempered flat target");
193 6 : keys.add("optional","TG_STRIDE","( default=1 ) number of AV_STRIDE between updates"
194 : " of target p(s) and reweighing factor c(t)");
195 : //optimization
196 6 : keys.add("compulsory","M_STEP","1.0","the mu step used for the Omega functional minimization");
197 6 : keys.add("compulsory","AV_STRIDE","500","number of simulation steps between alpha updates");
198 6 : keys.add("optional","TAU_MEAN","exponentially decaying average for alpha (rescaled using AV_STRIDE)."
199 : " Should be used only in very specific cases");
200 6 : keys.add("optional","INITIAL_ALPHA","( default=0 ) an initial guess for the bias potential parameter alpha");
201 6 : keys.addFlag("DAMPING_OFF",false,"do not use an AdaGrad-like term to rescale M_STEP");
202 : //output parameters file
203 6 : keys.add("compulsory","ALPHA_FILE","ALPHA","file name for output minimization parameters");
204 6 : keys.add("optional","PRINT_STRIDE","( default=10 ) stride for printing to ALPHA_FILE");
205 6 : keys.add("optional","FMT","specify format for ALPHA_FILE");
206 : //debug flags
207 6 : keys.addFlag("SERIAL",false,"perform the calculation in serial even if multiple tasks are available");
208 6 : keys.addFlag("MULTIPLE_WALKERS",false,"use multiple walkers connected via MPI for the optimization");
209 6 : keys.use("RESTART");
210 :
211 : //output components
212 12 : keys.addOutputComponent("rct","default","scalar","the reweighting factor c(t)");
213 12 : keys.addOutputComponent("work","default","scalar","the work done by the bias in one AV_STRIDE");
214 6 : }
215 :
216 4 : VesDeltaF::VesDeltaF(const ActionOptions&ao)
217 : : PLUMED_BIAS_INIT(ao)
218 4 : , isFirstStep_(true)
219 4 : , afterCalculate_(false)
220 4 : , mean_counter_(0)
221 4 : , av_counter_(0)
222 4 : , work_(0) {
223 : //set beta
224 4 : const double Kb=getKBoltzmann();
225 4 : double KbT=getkBT();
226 4 : plumed_massert(KbT>0,"your MD engine does not pass the temperature to plumed, you must specify it using TEMP");
227 4 : beta_=1.0/KbT;
228 :
229 : //initialize probability grids using local free energies
230 : bool spline=true;
231 : bool sparsegrid=false;
232 4 : std::string funcl="file.free"; //typical name given by sum_hills
233 :
234 : std::vector<std::string> fes_names;
235 8 : for(unsigned n=0;; n++) { //NB: here we start from FILE_F0 not from FILE_F1
236 : std::string filename;
237 24 : if(!parseNumbered("FILE_F",n,filename)) {
238 : break;
239 : }
240 8 : fes_names.push_back(filename);
241 8 : IFile gridfile;
242 8 : gridfile.open(filename);
243 8 : auto g=GridBase::create(funcl,getArguments(),gridfile,sparsegrid,spline,true);
244 : // we assume this cannot be sparse. in case we want it to be sparse, some of the methods
245 : // that are available only in Grid should be ported to GridBase
246 8 : auto gg=dynamic_cast<Grid*>(g.get());
247 : // if this throws, g is deleted
248 8 : plumed_assert(gg);
249 : // release ownership in order to transfer it to emplaced pointer
250 : // cppcheck-suppress ignoredReturnValue
251 : g.release();
252 8 : grid_p_.emplace_back(gg);
253 16 : }
254 4 : plumed_massert(grid_p_.size()>1,"at least 2 basins must be defined, starting from FILE_F0");
255 4 : alpha_size_=grid_p_.size()-1;
256 4 : sym_alpha_size_=alpha_size_*(alpha_size_+1)/2; //useful for symmetric matrix [alpha_size_]x[alpha_size_]
257 : //check for consistency with first local free energy
258 8 : for(unsigned n=1; n<grid_p_.size(); n++) {
259 8 : std::string error_tag="FILE_F"+std::to_string(n)+" '"+fes_names[n]+"' not compatible with reference one, FILE_F0";
260 4 : plumed_massert(grid_p_[n]->getSize()==grid_p_[0]->getSize(),error_tag);
261 4 : plumed_massert(grid_p_[n]->getMin()==grid_p_[0]->getMin(),error_tag);
262 4 : plumed_massert(grid_p_[n]->getMax()==grid_p_[0]->getMax(),error_tag);
263 4 : plumed_massert(grid_p_[n]->getBinVolume()==grid_p_[0]->getBinVolume(),error_tag);
264 : }
265 :
266 4 : bool no_mintozero=false;
267 4 : parseFlag("NO_MINTOZERO",no_mintozero);
268 4 : if(!no_mintozero) {
269 6 : for(unsigned n=0; n<grid_p_.size(); n++) {
270 4 : grid_p_[n]->setMinToZero();
271 : }
272 : }
273 4 : bool normalize=false;
274 4 : parseFlag("NORMALIZE",normalize);
275 4 : norm_.resize(grid_p_.size(),0);
276 4 : std::vector<double> c_norm(grid_p_.size());
277 : //convert the FESs to probability distributions
278 : //NB: the spline interpolation will be done on the probability distributions, not on the given FESs
279 : const unsigned ncv=getNumberOfArguments(); //just for ease
280 12 : for(unsigned n=0; n<grid_p_.size(); n++) {
281 808 : for(Grid::index_t t=0; t<grid_p_[n]->getSize(); t++) {
282 800 : std::vector<double> der(ncv);
283 800 : const double val=std::exp(-beta_*grid_p_[n]->getValueAndDerivatives(t,der));
284 1600 : for(unsigned s=0; s<ncv; s++) {
285 800 : der[s]*=-beta_*val;
286 : }
287 800 : grid_p_[n]->setValueAndDerivatives(t,val,der);
288 800 : norm_[n]+=val;
289 : }
290 8 : c_norm[n]=1./beta_*std::log(norm_[n]);
291 8 : if(normalize) {
292 4 : grid_p_[n]->scaleAllValuesAndDerivatives(1./norm_[n]);
293 4 : norm_[n]=1;
294 : }
295 : }
296 :
297 : //get target
298 4 : double biasfactor=0;
299 4 : parse("BIASFACTOR",biasfactor);
300 4 : plumed_massert(biasfactor==0 || biasfactor>1,"BIASFACTOR must be zero (for uniform target) or greater than one");
301 4 : if(biasfactor==0) {
302 2 : inv_gamma_=0;
303 : } else {
304 2 : inv_gamma_=1./biasfactor;
305 : }
306 4 : tg_counter_=0;
307 4 : tg_stride_=1;
308 4 : parse("TG_STRIDE",tg_stride_);
309 4 : tg_dV_dAlpha_.resize(alpha_size_,0);
310 4 : tg_d2V_dAlpha2_.resize(sym_alpha_size_,0);
311 :
312 : //setup optimization stuff
313 4 : minimization_step_=1;
314 4 : parse("M_STEP",minimization_step_);
315 :
316 4 : av_stride_=500;
317 4 : parse("AV_STRIDE",av_stride_);
318 4 : av_dV_dAlpha_.resize(alpha_size_,0);
319 4 : av_dV_dAlpha_prod_.resize(sym_alpha_size_,0);
320 4 : av_d2V_dAlpha2_.resize(sym_alpha_size_,0);
321 :
322 4 : mean_weight_tau_=0;
323 4 : parse("TAU_MEAN",mean_weight_tau_);
324 4 : if(mean_weight_tau_!=1) { //set it to 1 for basic SGD
325 4 : plumed_massert((mean_weight_tau_==0 || mean_weight_tau_>av_stride_),"TAU_MEAN is rescaled with AV_STRIDE, so it has to be greater");
326 4 : mean_weight_tau_/=av_stride_; //this way you can look at the number of simulation steps to choose TAU_MEAN
327 : }
328 :
329 8 : parseVector("INITIAL_ALPHA",mean_alpha_);
330 4 : if(mean_alpha_.size()>0) {
331 2 : plumed_massert(mean_alpha_.size()==alpha_size_,"provide one INITIAL_ALPHA for each basin beyond the first one");
332 : } else {
333 2 : mean_alpha_.resize(alpha_size_,0);
334 : }
335 4 : inst_alpha_=mean_alpha_;
336 4 : exp_alpha_.resize(alpha_size_);
337 8 : for(unsigned i=0; i<alpha_size_; i++) {
338 4 : exp_alpha_[i]=std::exp(-beta_*mean_alpha_[i]);
339 : }
340 4 : prev_exp_alpha_=exp_alpha_;
341 :
342 4 : damping_off_=false;
343 4 : parseFlag("DAMPING_OFF",damping_off_);
344 4 : if(damping_off_) {
345 2 : past_increment2_.resize(alpha_size_,1);
346 : } else {
347 2 : past_increment2_.resize(alpha_size_,0);
348 : }
349 :
350 : //file printing options
351 4 : std::string alphaFileName("ALPHA");
352 4 : parse("ALPHA_FILE",alphaFileName);
353 4 : print_stride_=10;
354 8 : parse("PRINT_STRIDE",print_stride_);
355 : std::string fmt;
356 4 : parse("FMT",fmt);
357 :
358 : //other flags, mainly for debugging
359 4 : NumParallel_=comm.Get_size();
360 4 : rank_=comm.Get_rank();
361 4 : bool serial=false;
362 4 : parseFlag("SERIAL",serial);
363 4 : if(serial) {
364 2 : log.printf(" -- SERIAL: running without loop parallelization\n");
365 2 : NumParallel_=1;
366 2 : rank_=0;
367 : }
368 :
369 4 : bool multiple_walkers=false;
370 4 : parseFlag("MULTIPLE_WALKERS",multiple_walkers);
371 4 : if(!multiple_walkers) {
372 2 : NumWalkers_=1;
373 : } else {
374 2 : if(comm.Get_rank()==0) { //multi_sim_comm works well on first rank only
375 2 : NumWalkers_=multi_sim_comm.Get_size();
376 : }
377 2 : if(comm.Get_size()>1) { //if each walker has more than one processor update them all
378 0 : comm.Bcast(NumWalkers_,0);
379 : }
380 : }
381 :
382 4 : checkRead();
383 :
384 : //restart if needed
385 4 : if(getRestart()) {
386 2 : IFile ifile;
387 2 : ifile.link(*this);
388 2 : if(NumWalkers_>1) {
389 4 : ifile.enforceSuffix("");
390 : }
391 2 : if(ifile.FileExist(alphaFileName)) {
392 2 : log.printf(" Restarting from: %s\n",alphaFileName.c_str());
393 2 : log.printf(" all options (also PRINT_STRIDE) must be consistent!\n");
394 2 : log.printf(" any INITIAL_ALPHA will be overwritten\n");
395 2 : ifile.open(alphaFileName);
396 : double time;
397 2 : std::vector<double> damping(alpha_size_);
398 20 : while(ifile.scanField("time",time)) { //room for improvements: only last line is important
399 16 : for(unsigned i=0; i<alpha_size_; i++) {
400 8 : const std::string index(std::to_string(i+1));
401 8 : prev_exp_alpha_[i]=std::exp(-beta_*mean_alpha_[i]);
402 16 : ifile.scanField("alpha_"+index,mean_alpha_[i]);
403 16 : ifile.scanField("auxiliary_"+index,inst_alpha_[i]);
404 16 : ifile.scanField("damping_"+index,damping[i]);
405 : }
406 8 : ifile.scanField();
407 8 : mean_counter_+=print_stride_;
408 : }
409 4 : for(unsigned i=0; i<alpha_size_; i++) {
410 2 : exp_alpha_[i]=std::exp(-beta_*mean_alpha_[i]);
411 2 : past_increment2_[i]=damping[i]*damping[i];
412 : }
413 : //sync all walkers and treads. Not sure is mandatory but is no harm
414 2 : comm.Barrier();
415 2 : if(comm.Get_rank()==0) {
416 2 : multi_sim_comm.Barrier();
417 : }
418 : } else {
419 0 : log.printf(" -- WARNING: restart requested, but no '%s' file found!\n",alphaFileName.c_str());
420 : }
421 2 : }
422 :
423 : //setup output file with Alpha values
424 4 : alphaOfile_.link(*this);
425 4 : if(NumWalkers_>1) {
426 2 : if(comm.Get_rank()==0 && multi_sim_comm.Get_rank()>0) {
427 : alphaFileName="/dev/null"; //only first walker writes on file
428 : }
429 4 : alphaOfile_.enforceSuffix("");
430 : }
431 4 : alphaOfile_.open(alphaFileName);
432 4 : if(fmt.length()>0) {
433 8 : alphaOfile_.fmtField(" "+fmt);
434 : }
435 :
436 : //add other output components
437 8 : addComponent("rct");
438 8 : componentIsNotPeriodic("rct");
439 8 : addComponent("work");
440 4 : componentIsNotPeriodic("work");
441 :
442 : //print some info
443 4 : log.printf(" Temperature T: %g\n",1./(Kb*beta_));
444 4 : log.printf(" Beta (1/Kb*T): %g\n",beta_);
445 4 : log.printf(" Local free energy basins files and normalization constants:\n");
446 12 : for(unsigned n=0; n<grid_p_.size(); n++) {
447 8 : log.printf(" F_%d filename: %s c_%d=%g\n",n,fes_names[n].c_str(),n,c_norm[n]);
448 : }
449 4 : if(no_mintozero) {
450 2 : log.printf(" -- NO_MINTOZERO: local free energies are not shifted to be zero at minimum\n");
451 : }
452 4 : if(normalize) {
453 2 : log.printf(" -- NORMALIZE: F_n+=c_n, alpha=DeltaF\n");
454 : }
455 4 : log.printf(" Using target distribution with 1/gamma = %g\n",inv_gamma_);
456 4 : log.printf(" and updated with stride %d\n",tg_stride_);
457 4 : log.printf(" Step for the minimization algorithm: %g\n",minimization_step_);
458 4 : log.printf(" Stride for the ensemble average: %d\n",av_stride_);
459 4 : if(mean_weight_tau_>1) {
460 2 : log.printf(" Exponentially decaying average with weight=tau/av_stride=%d\n",mean_weight_tau_);
461 : }
462 4 : if(mean_weight_tau_==1) {
463 0 : log.printf(" +++ WARNING +++ setting TAU_MEAN=1 is equivalent to use simple SGD, without mean alpha nor hessian contribution\n");
464 : }
465 4 : log.printf(" Initial guess for alpha:\n");
466 8 : for(unsigned i=0; i<alpha_size_; i++) {
467 4 : log.printf(" alpha_%d = %g\n",i+1,mean_alpha_[i]);
468 : }
469 4 : if(damping_off_) {
470 2 : log.printf(" -- DAMPING_OFF: the minimization step will NOT become smaller as the simulation goes on\n");
471 : }
472 4 : log.printf(" Printing on file %s with stride %d\n",alphaFileName.c_str(),print_stride_);
473 4 : if(serial) {
474 2 : log.printf(" -- SERIAL: running without loop parallelization\n");
475 : }
476 4 : if(NumParallel_>1) {
477 2 : log.printf(" Using multiple threads per simulation: %d\n",NumParallel_);
478 : }
479 4 : if(multiple_walkers) {
480 2 : log.printf(" -- MULTIPLE_WALKERS: multiple simulations will combine statistics for the optimization\n");
481 2 : if(NumWalkers_>1) {
482 2 : log.printf(" number of walkers: %d\n",NumWalkers_);
483 2 : log.printf(" walker rank: %d\n",multi_sim_comm.Get_rank()); //only comm.Get_rank()=0 prints, so this is fine
484 : } else {
485 0 : log.printf(" +++ WARNING +++ only one replica found: are you sure you are running MPI-connected simulations?\n");
486 : }
487 : }
488 4 : log.printf(" Bibliography ");
489 8 : log<<plumed.cite("Invernizzi and Parrinello, J. Chem. Theory Comput. 15, 2187-2194 (2019)");
490 8 : log<<plumed.cite("Valsson and Parrinello, Phys. Rev. Lett. 113, 090601 (2014)");
491 4 : if(inv_gamma_>0) {
492 4 : log<<plumed.cite("Valsson and Parrinello, J. Chem. Theory Comput. 11, 1996-2002 (2015)");
493 : }
494 :
495 : //last initializations
496 4 : prob_.resize(grid_p_.size());
497 4 : der_prob_.resize(grid_p_.size(),std::vector<double>(getNumberOfArguments()));
498 4 : update_tg_and_rct();
499 8 : }
500 :
501 804 : void VesDeltaF::calculate() {
502 : //get CVs
503 804 : const unsigned ncv=getNumberOfArguments(); //just for ease
504 804 : std::vector<double> cv(ncv);
505 1608 : for(unsigned s=0; s<ncv; s++) {
506 804 : cv[s]=getArgument(s);
507 : }
508 : //get probabilities for each basin, and total one
509 2412 : for(unsigned n=0; n<grid_p_.size(); n++) {
510 1608 : prob_[n]=grid_p_[n]->getValueAndDerivatives(cv,der_prob_[n]);
511 : }
512 804 : tot_prob_=prob_[0];
513 1608 : for(unsigned i=0; i<alpha_size_; i++) {
514 804 : tot_prob_+=prob_[i+1]*exp_alpha_[i];
515 : }
516 :
517 : //update bias and forces: V=-(1-inv_gamma_)*fes
518 804 : setBias((1-inv_gamma_)/beta_*std::log(tot_prob_));
519 1608 : for(unsigned s=0; s<ncv; s++) {
520 804 : double dProb_dCV_s=der_prob_[0][s];
521 1608 : for(unsigned i=0; i<alpha_size_; i++) {
522 804 : dProb_dCV_s+=der_prob_[i+1][s]*exp_alpha_[i];
523 : }
524 804 : setOutputForce(s,-(1-inv_gamma_)/beta_/tot_prob_*dProb_dCV_s);
525 : }
526 804 : afterCalculate_=true;
527 804 : }
528 :
529 804 : void VesDeltaF::update() {
530 : //skip first step to sync getTime() and av_counter_, as in METAD
531 804 : if(isFirstStep_) {
532 4 : isFirstStep_=false;
533 4 : return;
534 : }
535 800 : plumed_massert(afterCalculate_,"VesDeltaF::update() must be called after VesDeltaF::calculate() to work properly");
536 800 : afterCalculate_=false;
537 :
538 : //calculate derivatives for ensemble averages
539 800 : std::vector<double> dV_dAlpha(alpha_size_);
540 800 : std::vector<double> d2V_dAlpha2(sym_alpha_size_);
541 1600 : for(unsigned i=0; i<alpha_size_; i++) {
542 800 : dV_dAlpha[i]=-(1-inv_gamma_)/tot_prob_*prob_[i+1]*exp_alpha_[i];
543 : }
544 1600 : for(unsigned i=0; i<alpha_size_; i++) {
545 800 : d2V_dAlpha2[get_index(i,i)]=-beta_*dV_dAlpha[i];
546 1600 : for(unsigned j=i; j<alpha_size_; j++) {
547 800 : d2V_dAlpha2[get_index(i,j)]-=beta_/(1-inv_gamma_)*dV_dAlpha[i]*dV_dAlpha[j];
548 : }
549 : }
550 : //update ensemble averages
551 800 : av_counter_++;
552 1600 : for(unsigned i=0; i<alpha_size_; i++) {
553 800 : av_dV_dAlpha_[i]+=(dV_dAlpha[i]-av_dV_dAlpha_[i])/av_counter_;
554 1600 : for(unsigned j=i; j<alpha_size_; j++) {
555 800 : const unsigned ij=get_index(i,j);
556 800 : av_dV_dAlpha_prod_[ij]+=(dV_dAlpha[i]*dV_dAlpha[j]-av_dV_dAlpha_prod_[ij])/av_counter_;
557 800 : av_d2V_dAlpha2_[ij]+=(d2V_dAlpha2[ij]-av_d2V_dAlpha2_[ij])/av_counter_;
558 : }
559 : }
560 : //update work
561 800 : double prev_tot_prob=prob_[0];
562 1600 : for(unsigned i=0; i<alpha_size_; i++) {
563 800 : prev_tot_prob+=prob_[i+1]*prev_exp_alpha_[i];
564 : }
565 800 : work_+=(1-inv_gamma_)/beta_*std::log(tot_prob_/prev_tot_prob);
566 :
567 : //update coefficients
568 800 : if(av_counter_==av_stride_) {
569 16 : update_alpha();
570 16 : tg_counter_++;
571 16 : if(tg_counter_==tg_stride_) {
572 12 : update_tg_and_rct();
573 12 : tg_counter_=0;
574 : }
575 : //reset the ensemble averages
576 16 : av_counter_=0;
577 : std::fill(av_dV_dAlpha_.begin(),av_dV_dAlpha_.end(),0);
578 : std::fill(av_dV_dAlpha_prod_.begin(),av_dV_dAlpha_prod_.end(),0);
579 : std::fill(av_d2V_dAlpha2_.begin(),av_d2V_dAlpha2_.end(),0);
580 : }
581 : }
582 :
583 16 : void VesDeltaF::update_tg_and_rct() {
584 : //calculate target averages
585 16 : double Z_0=norm_[0];
586 32 : for(unsigned i=0; i<alpha_size_; i++) {
587 16 : Z_0+=norm_[i+1]*exp_alpha_[i];
588 : }
589 16 : double Z_tg=0;
590 : std::fill(tg_dV_dAlpha_.begin(),tg_dV_dAlpha_.end(),0);
591 : std::fill(tg_d2V_dAlpha2_.begin(),tg_d2V_dAlpha2_.end(),0);
592 1116 : for(Grid::index_t t=rank_; t<grid_p_[0]->getSize(); t+=NumParallel_) {
593 : //TODO can we recycle some code?
594 1100 : std::vector<double> prob(grid_p_.size());
595 3300 : for(unsigned n=0; n<grid_p_.size(); n++) {
596 2200 : prob[n]=grid_p_[n]->getValue(t);
597 : }
598 1100 : double tot_prob=prob[0];
599 2200 : for(unsigned i=0; i<alpha_size_; i++) {
600 1100 : tot_prob+=prob[i+1]*exp_alpha_[i];
601 : }
602 1100 : std::vector<double> dV_dAlpha(alpha_size_);
603 1100 : std::vector<double> d2V_dAlpha2(sym_alpha_size_);
604 2200 : for(unsigned i=0; i<alpha_size_; i++) {
605 1100 : dV_dAlpha[i]=-(1-inv_gamma_)/tot_prob*prob[i+1]*exp_alpha_[i];
606 : }
607 2200 : for(unsigned i=0; i<alpha_size_; i++) {
608 1100 : d2V_dAlpha2[get_index(i,i)]=-beta_*dV_dAlpha[i];
609 2200 : for(unsigned j=i; j<alpha_size_; j++) {
610 1100 : d2V_dAlpha2[get_index(i,j)]-=beta_/(1-inv_gamma_)*dV_dAlpha[i]*dV_dAlpha[j];
611 : }
612 : }
613 1100 : const double unnorm_tg_p=std::pow(tot_prob,inv_gamma_);
614 1100 : Z_tg+=unnorm_tg_p;
615 2200 : for(unsigned i=0; i<alpha_size_; i++) {
616 1100 : tg_dV_dAlpha_[i]+=unnorm_tg_p*dV_dAlpha[i];
617 : }
618 2200 : for(unsigned ij=0; ij<sym_alpha_size_; ij++) {
619 1100 : tg_d2V_dAlpha2_[ij]+=unnorm_tg_p*d2V_dAlpha2[ij];
620 : }
621 : }
622 16 : if(NumParallel_>1) {
623 10 : comm.Sum(Z_tg);
624 10 : comm.Sum(tg_dV_dAlpha_);
625 10 : comm.Sum(tg_d2V_dAlpha2_);
626 : }
627 32 : for(unsigned i=0; i<alpha_size_; i++) {
628 16 : tg_dV_dAlpha_[i]/=Z_tg;
629 : }
630 32 : for(unsigned ij=0; ij<sym_alpha_size_; ij++) {
631 16 : tg_d2V_dAlpha2_[ij]/=Z_tg;
632 : }
633 16 : getPntrToComponent("rct")->set(-1./beta_*std::log(Z_tg/Z_0)); //Z_tg is the best available estimate of Z_V
634 16 : }
635 :
636 16 : void VesDeltaF::update_alpha() {
637 : //combining the averages of multiple walkers
638 16 : if(NumWalkers_>1) {
639 8 : if(comm.Get_rank()==0) { //sum only once: in the first rank of each walker
640 8 : multi_sim_comm.Sum(av_dV_dAlpha_);
641 8 : multi_sim_comm.Sum(av_dV_dAlpha_prod_);
642 8 : multi_sim_comm.Sum(av_d2V_dAlpha2_);
643 16 : for(unsigned i=0; i<alpha_size_; i++) {
644 8 : av_dV_dAlpha_[i]/=NumWalkers_;
645 : }
646 16 : for(unsigned ij=0; ij<sym_alpha_size_; ij++) {
647 8 : av_dV_dAlpha_prod_[ij]/=NumWalkers_;
648 8 : av_d2V_dAlpha2_[ij]/=NumWalkers_;
649 : }
650 : }
651 8 : if(comm.Get_size()>1) { //if there are more ranks for each walker, everybody has to know
652 0 : comm.Bcast(av_dV_dAlpha_,0);
653 0 : comm.Bcast(av_dV_dAlpha_prod_,0);
654 0 : comm.Bcast(av_d2V_dAlpha2_,0);
655 : }
656 : }
657 : //set work and reset it
658 16 : getPntrToComponent("work")->set(work_);
659 16 : work_=0;
660 :
661 : //build the gradient and the Hessian of the functional
662 16 : std::vector<double> grad_omega(alpha_size_);
663 16 : std::vector<double> hess_omega(sym_alpha_size_);
664 32 : for(unsigned i=0; i<alpha_size_; i++) {
665 16 : grad_omega[i]=tg_dV_dAlpha_[i]-av_dV_dAlpha_[i];
666 32 : for(unsigned j=i; j<alpha_size_; j++) {
667 16 : const unsigned ij=get_index(i,j);
668 16 : hess_omega[ij]=beta_*(av_dV_dAlpha_prod_[ij]-av_dV_dAlpha_[i]*av_dV_dAlpha_[j])+tg_d2V_dAlpha2_[ij]-av_d2V_dAlpha2_[ij];
669 : }
670 : }
671 : //calculate the increment and update alpha
672 16 : mean_counter_++;
673 : long long unsigned mean_weight=mean_counter_;
674 16 : if(mean_weight_tau_>0 && mean_weight_tau_<mean_counter_) {
675 : mean_weight=mean_weight_tau_;
676 : }
677 16 : std::vector<double> damping(alpha_size_);
678 32 : for(unsigned i=0; i<alpha_size_; i++) {
679 16 : double increment_i=grad_omega[i];
680 32 : for(unsigned j=0; j<alpha_size_; j++) {
681 16 : increment_i+=hess_omega[get_index(i,j)]*(inst_alpha_[j]-mean_alpha_[j]);
682 : }
683 16 : if(!damping_off_) {
684 8 : past_increment2_[i]+=increment_i*increment_i;
685 : }
686 16 : damping[i]=std::sqrt(past_increment2_[i]);
687 16 : prev_exp_alpha_[i]=std::exp(-beta_*mean_alpha_[i]);
688 16 : inst_alpha_[i]-=minimization_step_/damping[i]*increment_i;
689 16 : mean_alpha_[i]+=(inst_alpha_[i]-mean_alpha_[i])/mean_weight;
690 16 : exp_alpha_[i]=std::exp(-beta_*mean_alpha_[i]);
691 : }
692 :
693 : //update the Alpha file
694 16 : if(mean_counter_%print_stride_==0) {
695 16 : alphaOfile_.printField("time",getTime());
696 32 : for(unsigned i=0; i<alpha_size_; i++) {
697 16 : const std::string index(std::to_string(i+1));
698 32 : alphaOfile_.printField("alpha_"+index,mean_alpha_[i]);
699 32 : alphaOfile_.printField("auxiliary_"+index,inst_alpha_[i]);
700 32 : alphaOfile_.printField("damping_"+index,damping[i]);
701 : }
702 16 : alphaOfile_.printField();
703 : }
704 16 : }
705 :
706 : //mapping of a [alpha_size_]x[alpha_size_] symmetric matrix into a vector of size sym_alpha_size_, useful for the communicator
707 4632 : inline unsigned VesDeltaF::get_index(const unsigned i, const unsigned j) const {
708 4632 : if(i<=j) {
709 4632 : return j+i*(alpha_size_-1)-i*(i-1)/2;
710 : } else {
711 0 : return get_index(j,i);
712 : }
713 : }
714 :
715 : }
716 : }
|
