-
Notifications
You must be signed in to change notification settings - Fork 3
/
mrpt.cpp
1957 lines (1721 loc) · 74.8 KB
/
mrpt.cpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. See the enclosed file LICENSE for a copy or if
* that was not distributed with this file, You can obtain one at
* http://mozilla.org/MPL/2.0/.
*
* Copyright 2017 Max H. Gerlach
*
* */
/*
* multireweighthisto-pt.cpp
*
* Created on: Jun 28, 2011
* Author: gerlach
*/
// generalized for SDW DQMC (2015-02-06 - )
#include <algorithm>
#include <omp.h>
#include <memory>
#include "boost/algorithm/string.hpp" // boost::split
#include "boost/filesystem.hpp"
#include "tools.h"
#include "mrpt.h"
#include "dataseriesloader.h"
#include "dataserieswriter.h"
#include "datamapwriter.h"
#include "statistics.h"
#include "numerics.h"
using namespace std;
template void printVector<LogVal>(const vector<LogVal>&);
template void printVector<double>(const vector<double>&);
MultireweightHistosPT::MultireweightHistosPT(ostream& outStream) :
numReplicas(0), systemN(0), systemSize(0),
minEnergyNormalized(0), maxEnergyNormalized(0), binCount(0), binSize(0), lBinSize(1.0),
out(outStream), basicConfig(false)
{
out << "max threads: " << omp_get_max_threads() << endl;
}
MultireweightHistosPT::~MultireweightHistosPT() {
}
void MultireweightHistosPT::addSimulationInfo(const std::string& filename) {
MetadataMap info = readOnlyMetadata(filename);
try {
// try to read info.dat as produced by detqmc
getMeta(info, "controlParameterName", controlParameterName);
std::string controlParameterValues_string; // "val1 val2 val3 [...]"
controlParameterValues_string = info["controlParameterValues"];
std::vector<std::string> controlParameterValues_string_vec;
boost::split(controlParameterValues_string_vec,
controlParameterValues_string,
boost::is_any_of(" "));
numReplicas = (uint32_t)controlParameterValues_string_vec.size();
controlParameterValues.clear();
controlParameterValues.reserve(numReplicas);
for (const auto& str : controlParameterValues_string_vec) {
controlParameterValues.push_back(fromString<double>(str));
}
} catch (KeyUndefined& exc) {
controlParameterName = "beta";
numReplicas = 0;
getMeta(info, "numTemperatures", numReplicas);
// betas should be stored in info.dat file below the metadata
DataSeriesLoader<double> dr;
dr.readFromFile(filename);
if (dr.getColumns() == 1) {
controlParameterValues = *dr.getData(0);
} else if (dr.getColumns() == 2) {
controlParameterValues = *dr.getData(1);
} else {
throw GeneralError("invalid format of inverse temperature file " + filename);
}
dr.deleteData();
}
systemN = 0;
getMeta(info, "N", systemN);
systemL = 0;
getMeta(info, "L", systemL);
infoNumSamples = 0;
std::string model;
getMeta(info, "model", model);
if (model == "sdw") {
// QMC simulation
unsigned timeslices;
getMeta(info, "m", timeslices);
double dtau = 0.1;
getMeta(info, "dtau", dtau);
systemSize = double(systemN) * double(timeslices) * dtau;
} else {
// classical MC simulation
systemSize = double(systemN);
}
try {
getMeta(info, "measureInterval", infoSweepsBetweenMeasurements);
} catch (KeyUndefined& exc) {
try {
getMeta(info, "sweepsBetweenMeasurements", infoSweepsBetweenMeasurements);
} catch (KeyUndefined& exc) {
infoSweepsBetweenMeasurements = 1;
}
}
// get a size hint for the time series
try {
int sweepsDone;
getMeta(info, "curSamples", sweepsDone);
infoNumSamples = sweepsDone / infoSweepsBetweenMeasurements;
} catch (KeyUndefined& exc) {
try {
getMeta(info, "curSamples", infoNumSamples);
} catch (KeyUndefined& exc) {
try {
getMeta(info, "totalSamples", infoNumSamples);
} catch (KeyUndefined& exc) {
try {
getMeta(info, "samples", infoNumSamples);
} catch (KeyUndefined& exc) {
try {
getMeta(info, "totalSweeps", infoNumSamples);
} catch (KeyUndefined& exc) {
infoNumSamples = 0;
}
}
}
}
}
out << "Simulation info loaded from " << filename << " numReplicas=" << numReplicas
<< " systemN=" << systemN << " systemSize=" << systemSize << std::endl;
addedEnergyTimeSeries = 0;
addedObservableTimeSeries = 0;
energyTimeSeries.resize(numReplicas);
observableTimeSeries.resize(numReplicas);
cpiTimeSeries.resize(numReplicas);
lZ_l.resize(numReplicas, LogVal(1.0));
out << "controlParameterName: " << controlParameterName << std::endl;
out << "controlParameterValues: ";
for (unsigned cpi = 0; cpi < controlParameterValues.size(); ++cpi) {
out << controlParameterValues[cpi] << " ";
}
out << std::endl;
basicConfig = true;
}
void MultireweightHistosPT::addInputTimeSeries_twoColumn(const std::string& filename, unsigned subsample, unsigned discardEntries) {
if (not basicConfig) {
throw GeneralError("tried to add input time series before basic configuration from simulation info was done (need to parse info.dat first)");
}
DoubleSeriesLoader input;
out << "adding time-series " << filename << " - " << flush;
input.readFromFile(filename, subsample, discardEntries, infoNumSamples);
unsigned replicaIndex;
input.getMeta("r", replicaIndex);
string obs;
input.getMeta("observable", obs);
out << obs << " - " << input.getData(0)->size() << endl;
if (obs == "energy") {
// if (energyTimeSeries.size() < replicaIndex + 1) {
// //content preserving resizes:
// unsigned newNumReplicas = replicaIndex + 1;
// energyTimeSeries.resize(newNumReplicas, 0);
// betaIndexTimeSeries.resize(newNumReplicas, 0);
// }
if (energyTimeSeries[replicaIndex] != 0) {
throw GeneralError("two time series added for " + obs + " from replica-index " + numToString(replicaIndex));
}
energyTimeSeries[replicaIndex] = input.getData(input.getColumns() - 1);
if (input.getColumns() != 2) {
throw GeneralError(filename + ": expected a two column time series file, but got " + numToString(input.getColumns()) + " column(s)");
}
++addedEnergyTimeSeries;
//update cpiTimeSeries:
std::shared_ptr<std::vector<double>> dCPItimeSeries = input.getData(0); //values loaded in as doubles...
cpiTimeSeries[replicaIndex] = std::shared_ptr<std::vector<int>>(
new vector<int>(energyTimeSeries[replicaIndex]->size()));
for (unsigned n = 0; n < dCPItimeSeries->size(); ++n) {
int cpi = static_cast<int>((*dCPItimeSeries)[n]); //TODO:useless cast!
(*(cpiTimeSeries[replicaIndex]))[n] = cpi;
}
// delete dCPItimeSeries; //don't need this in memory anymore
} else if (observable == "" or obs == observable) {
observable = obs;
// if (observableTimeSeries.size() < replicaIndex + 1) {
// observableTimeSeries.resize(replicaIndex + 1, 0);
// }
if (observableTimeSeries[replicaIndex] != 0) {
throw GeneralError("two time series added for " + obs + " from replica-index " + numToString(replicaIndex));
}
observableTimeSeries[replicaIndex] = input.getData(input.getColumns() - 1);
++addedObservableTimeSeries;
if (input.getColumns() == 2) {
input.getData(0)->clear(); //don't need this in memory
}
} else {
throw GeneralError("in " + filename + ": Observable doesn't match previous\n" +
obs + " vs. " + observable + "\n");
}
unsigned N;
input.getMeta("N", N);
if (systemN == 0) {
systemN = N;
} else if (systemN != N) {
throw GeneralError("in " + filename + ": Non matching system sizes: " +
numToString(N) + " vs. " + numToString(systemN));
}
}
void MultireweightHistosPT::addInputTimeSeries_singleColumn(const std::string& filename, unsigned subsample, unsigned discardEntries) {
if (not basicConfig) {
throw GeneralError("tried to add input time series before basic configuration from simulation info was done (need to parse info.dat first)");
}
DoubleSeriesLoader input;
out << "adding time-series " << filename << " - " << flush;
input.readFromFile(filename, subsample, discardEntries, infoNumSamples);
if (input.getColumns() != 1) {
throw GeneralError(filename + ": expected a single column time series file, but got " +
numToString(input.getColumns()) + " columns");
}
unsigned replicaIndex;
try {
input.getMeta("replicaIndex", replicaIndex);
} catch (KeyUndefined& exc) {
try {
input.getMeta("controlParameterIndex", replicaIndex);
} catch (KeyUndefined& exc) {
// controlParameterIndex not defined --> get it from the actual controlParameter value
double targetValue;
input.getMeta(controlParameterName, targetValue);
replicaIndex = (uint32_t)findNearest(controlParameterValues, targetValue);
}
}
string obs;
input.getMeta("observable", obs);
out << obs << " - " << input.getData(0)->size() << endl;
if (obs == "energy" or obs == "associatedEnergy") {
// if (energyTimeSeries.size() < replicaIndex + 1) {
// //content preserving resizes:
// unsigned newNumReplicas = replicaIndex + 1;
// energyTimeSeries.resize(newNumReplicas, 0);
// betaIndexTimeSeries.resize(newNumReplicas, 0);
// }
if (energyTimeSeries[replicaIndex] != 0) {
throw GeneralError("two time series added for " + obs +
" from replica-index " + numToString(replicaIndex));
}
energyTimeSeries[replicaIndex] = input.getData();
++addedEnergyTimeSeries;
//update cpiTimeSeries -- just fill with constant controlparameter index
if (cpiTimeSeries[replicaIndex] == 0) {
cpiTimeSeries[replicaIndex] = std::shared_ptr<vector<int>>(
new vector<int>(energyTimeSeries[replicaIndex]->size(), replicaIndex));
}
} else if (observable == "" or obs == observable) {
observable = obs;
// if (observableTimeSeries.size() < replicaIndex + 1) {
// observableTimeSeries.resize(replicaIndex + 1, 0);
// }
if (observableTimeSeries[replicaIndex] != 0) {
throw GeneralError("two time series added for " + obs +
" from replica-index " + numToString(replicaIndex));
}
observableTimeSeries[replicaIndex] = input.getData();
++addedObservableTimeSeries;
//update cpiTimeSeries -- just fill with constant controlparameter index
if (cpiTimeSeries[replicaIndex] == 0) {
cpiTimeSeries[replicaIndex] = std::shared_ptr<vector<int>>(
new vector<int>(observableTimeSeries[replicaIndex]->size(), replicaIndex));
}
} else {
throw GeneralError("in " + filename + ": Observable doesn't match previous\n" +
obs + " vs. " + observable + "\n");
}
unsigned N;
input.getMeta("N", N);
if (systemN == 0) {
systemN = N;
} else if (systemN != N) {
throw GeneralError("in " + filename + ": Non matching system sizes: " +
numToString(N) + " vs. " + numToString(systemN));
}
}
void MultireweightHistosPT::sortTimeSeriesByControlParameter() {
out << "Sorting time series by control parameter..." << flush;
if (addedObservableTimeSeries > 0) {
//all time series need to have the same length
unsigned M = (unsigned)energyTimeSeries[0]->size();
for (unsigned r = 1; r < numReplicas; ++r) {
if (energyTimeSeries[r]->size() != M or
observableTimeSeries[r]->size() != M) {
throw GeneralError("Time series length mismatch: for control-parameter-sorted "
"timeseries all need to have the same length");
}
}
//for each sample:
//first these vectors hold the original ordering by replica index
//then they are sorted back into the time series, sorted by beta index
vector<double> tempEnergies(numReplicas);
vector<double> tempObs(numReplicas);
vector<int> tempCPI(numReplicas);
for (unsigned m = 0; m < M; ++m) {
for (unsigned r = 0; r < numReplicas; ++r) {
tempEnergies[r] = (*energyTimeSeries[r])[m];
tempObs[r] = (*observableTimeSeries[r])[m];
tempCPI[r] = (*cpiTimeSeries[r])[m];
}
for (unsigned r = 0; r < numReplicas; ++r) {
int bi = tempCPI[r];
(*energyTimeSeries[bi])[m] = tempEnergies[r];
(*observableTimeSeries[bi])[m] = tempObs[r];
//the following is now trivial, but later code depends on it
(*cpiTimeSeries[bi])[m] = bi;
}
}
} else {
//all time series need to have the same length
unsigned M = (unsigned)energyTimeSeries[0]->size();
for (unsigned r = 1; r < numReplicas; ++r) {
if (energyTimeSeries[r]->size() != M) {
throw GeneralError("Time series length mismatch: for control-parameter-sorted "
"timeseries all need to have the same length");
}
}
//for each sample:
//first these vectors hold the original ordering by replica index
//then they are sorted back into the time series, sorted by beta index
vector<double> tempEnergies(numReplicas);
vector<int> tempCPI(numReplicas);
for (unsigned m = 0; m < M; ++m) {
for (unsigned r = 0; r < numReplicas; ++r) {
tempEnergies[r] = (*energyTimeSeries[r])[m];
tempCPI[r] = (*cpiTimeSeries[r])[m];
}
for (unsigned r = 0; r < numReplicas; ++r) {
int bi = tempCPI[r];
(*energyTimeSeries[bi])[m] = tempEnergies[r];
//the following is now trivial, but later code depends on it
(*cpiTimeSeries[bi])[m] = bi;
}
}
}
out << endl;
}
MultireweightHistosPT::ResultsMap* MultireweightHistosPT::
directNoReweighting() {
out << "Computing estimates from time series without any reweighting... "
<< flush;
ResultsMap* results = new ResultsMap;
//vectors betaIndex (of this->betas) -> value at that temperature
//everything normalized by system volume
vector<double> meanEnergy_l(numReplicas, 0);
vector<double> meanEnergySquared_l(numReplicas, 0);
vector<double> meanObs_l(numReplicas, 0);
vector<double> meanObsSquared_l(numReplicas);
vector<double> meanObsToTheFourth_l(numReplicas, 0);
//calculate for each replica separately:
#pragma omp parallel for
for (signed k = 0; k < (signed)numReplicas; ++k) {
//vectors for this replica (still map betaindex -> values):
vector<double> k_meanEnergy_l(numReplicas, 0);
vector<double> k_meanEnergySquared_l(numReplicas, 0);
vector<double> k_meanObs_l(numReplicas, 0);
vector<double> k_meanObsSquared_l(numReplicas);
vector<double> k_meanObsToTheFourth_l(numReplicas, 0);
unsigned N_k = (unsigned)energyTimeSeries[k]->size();
//sum up:
for (unsigned n = 0; n < N_k; ++n) {
int cpi = (*cpiTimeSeries[k])[n];
double e = (*energyTimeSeries[k])[n];
double o = (*observableTimeSeries[k])[n];
k_meanEnergy_l[cpi] += e;
k_meanEnergySquared_l[cpi] += e*e;
k_meanObs_l[cpi] += o;
k_meanObsSquared_l[cpi] += o*o;
k_meanObsToTheFourth_l[cpi] += o*o*o*o;
}
//divide to form averages, divisor: no of samples for this temperature
//in the time series (numReplicas == #temperatures)
double divisor = double(N_k) / double(numReplicas);
for (unsigned cpi = 0; cpi < numReplicas; ++cpi) {
k_meanEnergy_l[cpi] /= divisor;
k_meanEnergySquared_l[cpi] /= divisor;
k_meanObs_l[cpi] /= divisor;
k_meanObsSquared_l[cpi] /= divisor;
k_meanObsToTheFourth_l[cpi] /= divisor;
}
//average over replicas -- sum up everything from all replicas
#pragma omp critical
{
for (unsigned cpi = 0; cpi < numReplicas; ++cpi) {
meanEnergy_l[cpi] += k_meanEnergy_l[cpi];
meanEnergySquared_l[cpi] += k_meanEnergySquared_l[cpi];
meanObs_l[cpi] += k_meanObs_l[cpi];
meanObsSquared_l[cpi] += k_meanObsSquared_l[cpi];
meanObsToTheFourth_l[cpi] +=
k_meanObsToTheFourth_l[cpi];
}
}
}
//average over replicas -- divide
for (unsigned cpi = 0; cpi < numReplicas; ++cpi) {
meanEnergy_l[cpi] /= double(numReplicas);
meanEnergySquared_l[cpi] /= double(numReplicas);
meanObs_l[cpi] /= double(numReplicas);
meanObsSquared_l[cpi] /= double(numReplicas);
meanObsToTheFourth_l[cpi] /= double(numReplicas);
}
//compute final results and put them into the map
for (unsigned cpi = 0; cpi < numReplicas; ++cpi) {
double cp = controlParameterValues[cpi];
double heatCapacity = systemSize * cp*cp *
(meanEnergySquared_l[cpi] - pow(meanEnergy_l[cpi], 2));
double suscPartObs = systemSize * meanObsSquared_l[cpi];
double suscObs = systemSize *
(meanObsSquared_l[cpi] - pow(meanObs_l[cpi], 2));
double binderObs = 1.0 - (meanObsToTheFourth_l[cpi] /
(3 * pow(meanObsSquared_l[cpi], 2)));
double binderRatioObs = (meanObsToTheFourth_l[cpi] /
(pow(meanObsSquared_l[cpi], 2)));
//set results without specification of errors:
(*results)[cp] = ReweightingResult(meanEnergy_l[cpi],
heatCapacity,
meanObs_l[cpi],
suscPartObs,
suscObs,
binderObs,
binderRatioObs);
}
out << "done." << endl;
return results;
}
void MultireweightHistosPT::setUpHistograms(int binCount_) {
if (addedEnergyTimeSeries == 0) {
throw GeneralError("No energy time series added!");
}
if (addedObservableTimeSeries == 0) {
cout << "Warning: No observable time series added." << endl;
}
if (numReplicas != controlParameterValues.size() or numReplicas != addedEnergyTimeSeries
or (addedObservableTimeSeries > 0 and
numReplicas != addedObservableTimeSeries)) {
// throw GeneralError("Mismatched number of replicas (forgot to add time series? wrong betas?)");
cout << "Warning: Number of added time series does not match number "
"of replicas from simulation info file." << endl;
}
//replace null-pointer with zero-length vectors at the spots, where
//no time series were added
for (unsigned k = 0; k < numReplicas; ++k) {
if (not energyTimeSeries[k]) {
energyTimeSeries[k].reset(new vector<double>(0));
}
if (not observableTimeSeries[k]) {
observableTimeSeries[k].reset(new vector<double>(0));
}
if (not cpiTimeSeries[k]) {
cpiTimeSeries[k].reset(new vector<int>(0));
}
}
binCount = binCount_;
findMinMax(energyTimeSeries, minEnergyNormalized, maxEnergyNormalized);
minEnergy = minEnergyNormalized * systemSize;
maxEnergy = maxEnergyNormalized * systemSize;
deltaU = (maxEnergyNormalized - minEnergyNormalized) / binCount;
const double SMALL = 1e-10; //to fit maxEnergy into the highest bin
binSize = (maxEnergyNormalized - minEnergyNormalized + SMALL) / binCount;
lBinSize = LogVal(binSize);
if (addedObservableTimeSeries > 0) {
findMinMax(observableTimeSeries, minObservableNormalized, maxObservableNormalized);
}
U_m.resize(binCount);
for (unsigned m = 0; m < binCount; ++m) {
//take normalized energy at bin center, then undo normalization
U_m[m] = (minEnergyNormalized + binSize * (m + 0.5)) * systemSize;
}
lZ_l = vector<LogVal>(numReplicas, LogVal(1.0));
H_km.resize(boost::extents[numReplicas][binCount]);
initArray(H_km, 0);
g_km.resize(boost::extents[numReplicas][binCount]);
initArray(g_km, 0);
H_lm.resize(boost::extents[numReplicas][binCount]);
initArray(H_lm, 0);
N_kl.resize(boost::extents[numReplicas][numReplicas]);
initArray(N_kl, 0);
H_m = Histo(binCount, 0);
m_kn = IntSeriesCollection(numReplicas);
for (unsigned k = 0; k < numReplicas; ++k) {
m_kn[k].reset(new std::vector<int>(energyTimeSeries[k]->size()));
}
lOmega_m.resize(binCount, LogVal(1.0));
}
void MultireweightHistosPT::setUpHistogramsIsing() {
if (addedEnergyTimeSeries == 0) {
throw GeneralError("No energy time series added!");
}
if (addedObservableTimeSeries == 0) {
cout << "Warning: No observable time series added." << endl;
}
if (numReplicas != controlParameterValues.size() or numReplicas != addedEnergyTimeSeries
or (addedObservableTimeSeries > 0 and
numReplicas != addedObservableTimeSeries)) {
// throw GeneralError("Mismatched number of replicas (forgot to add time series? wrong betas?)");
cout << "Warning: Number of added time series does not match number "
"of replicas from simulation info file." << endl;
}
//replace null-pointer with zero-length vectors at the spots, where
//no time series were added
for (unsigned k = 0; k < numReplicas; ++k) {
if (not energyTimeSeries[k]) {
energyTimeSeries[k].reset(new vector<double>(0));
}
if (not observableTimeSeries[k]) {
observableTimeSeries[k].reset(new vector<double>(0));
}
if (not cpiTimeSeries[k]) {
cpiTimeSeries[k].reset(new vector<int>(0));
}
}
findMinMax(energyTimeSeries, minEnergyNormalized, maxEnergyNormalized);
//actual min and max energies:
minEnergy = round(minEnergyNormalized * double(systemSize));
maxEnergy = round(maxEnergyNormalized * double(systemSize));
//special knowledge of 2D Ising model:
//offset by half a bin width, else the binning routines won't work due to
//rounding errors:
// minEnergyNormalized -= double(2) / systemSize;
//the energies -2N+4 and 2N-4 aren't physically accessible, but we'll
//include them in the histograms anyway
binCount = int(maxEnergy - minEnergy) / 4 + 1;
deltaU = 4.0;
const double SMALL = 1e-10; //to fit maxEnergy into the highest bin
binSize = deltaU / double(systemSize) + SMALL;
lBinSize = LogVal(binSize);
if (addedObservableTimeSeries) {
findMinMax(observableTimeSeries, minObservableNormalized, maxObservableNormalized);
}
U_m.resize(binCount);
for (int m = 0; m < int(binCount); ++m) {
U_m[m] = double(minEnergy) + m * deltaU;
}
lZ_l = vector<LogVal>(numReplicas, LogVal(1.0));
H_km.resize(boost::extents[numReplicas][binCount]);
initArray(H_km, 0);
g_km.resize(boost::extents[numReplicas][binCount]);
initArray(g_km, 0);
H_lm.resize(boost::extents[numReplicas][binCount]);
initArray(H_lm, 0);
N_kl.resize(boost::extents[numReplicas][numReplicas]);
initArray(N_kl, 0);
H_m = Histo(binCount, 0);
m_kn = IntSeriesCollection(numReplicas);
for (unsigned k = 0; k < numReplicas; ++k) {
m_kn[k].reset(new std::vector<int>(energyTimeSeries[k]->size()));
}
lOmega_m.resize(binCount, LogVal(1.0));
}
void MultireweightHistosPT::createHistogramsHelper() {
out << "Creating energy histograms etc. minEnergyNormalized=" << minEnergyNormalized
<< " maxEnergyNormalized=" << maxEnergyNormalized
<< " binSize=" << binSize
<< " binCount=" << binCount << endl;
// #pragma omp parallel for
for (int k = 0; k < (signed)numReplicas; ++k) {
for (unsigned n = 0; n < energyTimeSeries[k]->size(); ++n) {
//get bin number, correct due to rounding in cast:
int m = static_cast<int>(((*energyTimeSeries[k])[n] - minEnergyNormalized) / binSize);
//get beta index
int l = (*(cpiTimeSeries[k]))[n];
N_kl[k][l] += 1;
(*m_kn[k])[n] = m;
++H_km[k][m];
// #pragma omp atomic
++H_m[m];
// #pragma omp atomic
++H_lm[l][m];
}
out << "." << flush;
}
//not needed any more
cpiTimeSeries.clear();
out << " done" << endl;
}
void MultireweightHistosPT::createHistogramsHelperDiscrete() {
out << "Creating energy histograms etc. minEnergyNormalized=" << minEnergyNormalized
<< " maxEnergyNormalized=" << maxEnergyNormalized
<< " binSize=" << binSize
<< " binCount=" << binCount
<< " for originally discrete bins " << endl;
// #pragma omp parallel for
for (int k = 0; k < (signed)numReplicas; ++k) {
for (unsigned n = 0; n < energyTimeSeries[k]->size(); ++n) {
//get bin number, correct due to rounding in cast:
// int m = static_cast<int>(((*energyTimeSeries[k])[n] - minEnergyNormalized) / binSize);
int m = int(round(((*energyTimeSeries[k])[n] * systemSize - minEnergy) / deltaU));
//get beta index
int l = (*(cpiTimeSeries[k]))[n];
N_kl[k][l] += 1;
(*m_kn[k])[n] = m;
++H_km[k][m];
// #pragma omp atomic
++H_m[m];
// #pragma omp atomic
++H_lm[l][m];
}
out << "." << flush;
}
//not needed any more
cpiTimeSeries.clear();
// destroyAll(cpiTimeSeries);
out << " done" << endl;
}
void MultireweightHistosPT::createHistograms(int binCount_) {
setUpHistograms(binCount_);
createHistogramsHelper();
}
void MultireweightHistosPT::createHistogramsIsing() {
setUpHistogramsIsing();
createHistogramsHelperDiscrete();
}
void MultireweightHistosPT::measureGlobalInefficiencies(bool saveAutocorr) {
out << "Measuring energy time series inefficiencies g_k, "
"ignoring differences in individual bins" << endl;
g_km.resize(boost::extents[numReplicas][binCount]);
namespace fs = boost::filesystem;
if (saveAutocorr) {
fs::path autocorrPath("./autocorr");
fs::create_directory(autocorrPath);
fs::current_path(autocorrPath);
}
#pragma omp parallel for
for (int k = 0; k < (signed)numReplicas; ++k) {
std::shared_ptr<AutoCorrMap> autocorr;
if (saveAutocorr) {
autocorr.reset(new AutoCorrMap);
}
double tauint = tauint_adaptive(energyTimeSeries[k].get(), autocorr.get());
double g_k = 1 + 2 * tauint;
for (unsigned m = 0; m < binCount; ++m) {
g_km[k][m] = g_k;
}
if (saveAutocorr) {
IntDoubleMapWriter writeAutocorr;
writeAutocorr.setData(autocorr);
writeAutocorr.addHeaderText("Autocorrelation function of the energy");
writeAutocorr.addHeaderText("Replica-index (corresponds to control-parameter-index if time series were sorted):");
writeAutocorr.addMeta("k", k);
writeAutocorr.addHeaderText("Estimated integrated autocorrelation time");
writeAutocorr.addMeta("tau-int", tauint);
writeAutocorr.addMeta("g", g_k);
writeAutocorr.addHeaderText("gap t \t autocorr");
writeAutocorr.writeToFile("autocorr-k" + numToString(k) + ".dat");
}
out << "." << flush;
}
if (saveAutocorr) {
fs::current_path("..");
}
out << " done" << endl;
}
void MultireweightHistosPT::writeOutEnergyTauInt(const std::string& filename) {
out << "Estimating energy tau-ints... " << endl;
std::shared_ptr<map<double, double> > tauint(new map<double, double>);
#pragma omp parallel for
for (int k = 0; k < (signed)numReplicas; ++k) {
tauint->insert(make_pair(controlParameterValues[k],
tauint_adaptive(energyTimeSeries[k].get()) *
infoSweepsBetweenMeasurements));
cout << "." << flush;
}
DoubleMapWriter w;
w.addHeaderText("Integrated autocorrelation time of the energy, used fast adaptive method");
w.addHeaderText("Unit: Monte Carlo Sweeps (even if less measured samples were taken)");
w.addMeta("sweepsBetweenMeasurements", infoSweepsBetweenMeasurements);
w.addMeta("observable", "energy");
w.addMeta("L", systemL);
w.addMeta("N", systemN);
w.addMeta("systemSize", systemSize);
w.addMeta("controlParameterName", controlParameterName);
w.addHeaderText("cp \t tau-int");
w.setData(tauint);
w.writeToFile(filename);
out << " Done." << endl;
}
void MultireweightHistosPT::writeOutObsTauInt(const std::string& filename) {
out << "Estimating " << observable << " tau-ints... " << endl;
std::shared_ptr<map<double, double> > tauint(new map<double, double>);
#pragma omp parallel for
for (int k = 0; k < (signed)numReplicas; ++k) {
tauint->insert(make_pair(controlParameterValues[k],
tauint_adaptive(observableTimeSeries[k].get()) *
infoSweepsBetweenMeasurements));
cout << "." << flush;
}
DoubleMapWriter w;
w.addHeaderText("Integrated autocorrelation time of " + observable +
", used fast adaptive method");
w.addHeaderText("Unit: Monte Carlo Sweeps (even if less measured samples were taken)");
w.addMeta("sweepsBetweenMeasurements", infoSweepsBetweenMeasurements);
w.addMeta("observable", observable);
w.addMeta("L", systemL);
w.addMeta("N", systemN);
w.addMeta("systemSize", systemSize);
w.addMeta("controlParameterName", controlParameterName);
w.addHeaderText("cp \t tau-int");
w.setData(tauint);
w.writeToFile(filename);
out << " Done." << endl;
}
void MultireweightHistosPT::setBinInefficienciesToUnity() {
out << "Setting all bin inefficiencies g_km to 1";
g_km.resize(boost::extents[numReplicas][binCount]);
std::fill(g_km.data(), g_km.data() + g_km.num_elements(), 1.0);
out << endl;
}
void MultireweightHistosPT::measureBinInefficiencies(bool saveAutocorr) {
out << "Measuring bin inefficiencies g_km" << endl;
g_km.resize(boost::extents[numReplicas][binCount]);
std::fill(g_km.data(), g_km.data() + g_km.num_elements(), 1.0);
namespace fs = boost::filesystem;
if (saveAutocorr) {
fs::path autocorrPath("./autocorr");
fs::create_directory(autocorrPath);
fs::current_path(autocorrPath);
for (unsigned k = 0; k < numReplicas; ++k) {
if (energyTimeSeries[k]->size() > 0) {
fs::path dir("k" + numToString(k));
fs::create_directory(dir);
}
}
}
#pragma omp parallel for
for (int k = 0; k < (signed)numReplicas; ++k) {
vector<std::shared_ptr<AutoCorrMap> > autocorr_m(binCount, 0);
if (saveAutocorr) {
for (unsigned m = 0; m < binCount; ++m) {
autocorr_m[m].reset(new AutoCorrMap);
}
}
//adaptive integration of autocorrelation function
//high time resolution for small lag times, lower resolution for higher times in the
//slowly decaying tail of the autocorrelation function
vector<char> zeroCrossed(binCount, false);
unsigned countZeroCrossed = 0;
unsigned N = (unsigned)energyTimeSeries[k]->size();
if (N == 0) {
//skip empty time series
continue;
}
vector<double> binMeanOccupation(binCount, 0);
vector<double> binMeanSquared(binCount, 0);
for (unsigned m = 0; m < binCount; ++m) {
binMeanOccupation[m] = double(H_km[k][m]) / double(N);
binMeanSquared[m] = pow(binMeanOccupation[m], 2);
if (binMeanOccupation[m] == 0) {
//bin never occupied in this replica --> g_km has no real sense
//==> don't take this bin into account at all in the following, leave g_km at 1.0
zeroCrossed[m] = true;
++countZeroCrossed;
}
}
unsigned i = 1;
unsigned t_i = 1;
do {
//compute bin autocorrelation functions
vector<double> binAutoCorr(binCount, 0);
for (unsigned n = 0; n < N - t_i; ++n) {
int bin1 = (*m_kn[k])[n];
int bin2 = (*m_kn[k])[n + t_i];
if (bin1 == bin2) {
binAutoCorr[bin1] += 1.0;
}
}
for (unsigned m = 0; m < binCount; ++m) {
if (not zeroCrossed[m]) {
binAutoCorr[m] /= double(N - t_i);
binAutoCorr[m] -= binMeanSquared[m];
binAutoCorr[m] /= (binMeanOccupation[m] - binMeanSquared[m]); //normalize by variance
if (binAutoCorr[m] <= 0) {
zeroCrossed[m] = true;
++countZeroCrossed;
} else {
//weighted addition to estimate integrated autocorrelation time
double t_next = 1.0 + (i + 1.0) * (i) / 2.0;
g_km[k][m] += 2.0 * binAutoCorr[m] * (t_next - t_i);
if (saveAutocorr) {
autocorr_m[m]->insert(make_pair(t_i, binAutoCorr[m]));
}
}
}
}
i += 1;
//lag time for next step of the iteration
t_i = unsigned(1.0 + i * (i - 1.0) / 2.0);
} while (t_i < N - 1 and countZeroCrossed < binCount);
if (saveAutocorr) {
for (unsigned m = 0; m < binCount; ++m) {
IntDoubleMapWriter wa;
wa.setData(autocorr_m[m]);
wa.addHeaderText("Autocorrelation function of the characteristic function of one energy bin");
wa.addHeaderText("Replica-index (corresponds to control-parameter-index if time series were sorted):");
wa.addMeta("k", k);
wa.addHeaderText("Bin-index");
wa.addMeta("m", m);
wa.addHeaderText("Corresponding energy");
wa.addMeta("U_m", U_m[m]);
wa.addHeaderText("Corresponding energy normalized by systemSize");
wa.addMeta("e_m", double(U_m[m]) / double(systemSize));
wa.addHeaderText("Statistical inefficiency estimated from integrated autocorrelation time");
wa.addMeta("g_km", g_km[k][m]);
wa.addHeaderText("gap t \t autocorr");
wa.writeToFile("k" + numToString(k) + "/autocorr-k" + numToString(k)
+ "m-" + numToString(m) + ".dat");
}
}
out << "." << flush;
}
if (saveAutocorr) {
fs::current_path("..");
}
out << " done" << endl;
}
void MultireweightHistosPT::saveg_km(const std::string& filename) {
ofstream outgkm(filename.c_str());
for (unsigned k = 0; k < numReplicas; ++k) {
for (unsigned m = 0; m < binCount; ++m) {
outgkm << g_km[k][m] << '\t';
}
outgkm << '\n';
}
}
void MultireweightHistosPT::saveH_km(const std::string& filename) {
ofstream outHkm(filename.c_str());
for (unsigned k = 0; k < numReplicas; ++k) {
for (unsigned m = 0; m < binCount; ++m) {
outHkm << H_km[k][m] << '\t';
}
outHkm << '\n';
}
}
void MultireweightHistosPT::saveU_m(const std::string& filename) {
ofstream outUm(filename.c_str());
for (unsigned m = 0; m < binCount; ++m) {
outUm << U_m[m] << '\n';
}
}
void MultireweightHistosPT::saveNeff_lm(const std::string& filename) {
ofstream output(filename.c_str());
for (unsigned l = 0; l < numReplicas; ++l) {
for (unsigned m = 0; m < binCount; ++m) {
output << Neff_lm[l][m] << '\t';
}
output << '\n';
}
}
void MultireweightHistosPT::saveNeff_l(const std::string& filename) {
ofstream output(filename.c_str());
for (unsigned l = 0; l < numReplicas; ++l) {
output << Neff_l[l] << '\n';
}
}
void MultireweightHistosPT::savelNeff_lm(const std::string& filename) {
ofstream output(filename.c_str());
for (unsigned l = 0; l < numReplicas; ++l) {
for (unsigned m = 0; m < binCount; ++m) {
output << lNeff_lm[l][m] << '\t';
}
output << '\n';
}
}
inline void MultireweightHistosPT::updateEffectiveCounts() {
Heff_m.resize(binCount, 0);
Neff_lm.resize(boost::extents[numReplicas][binCount]);
initArray(Neff_lm, 0);
for (unsigned k = 0; k < numReplicas; ++k) {
for (unsigned m = 0; m < binCount; ++m) {
Heff_m[m] += double(H_km[k][m]) / g_km[k][m];
for (unsigned l = 0; l < numReplicas; ++l) {
Neff_lm[l][m] += double(N_kl[k][l]) / g_km[k][m];
}
}
}
lHeff_m.resize(binCount);
lNeff_lm.resize(boost::extents[numReplicas][binCount]);
lPrecalc_lm.resize(boost::extents[numReplicas][binCount]);
Neff_l.resize(numReplicas, 0);
for (unsigned m = 0; m < binCount; ++m) {
lHeff_m[m] = (Heff_m[m] != 0 ? LogVal(Heff_m[m]) : LogVal(LogVal::LogZero));
for (unsigned l = 0; l < numReplicas; ++l) {
lNeff_lm[l][m] = (Neff_lm[l][m] != 0 ? LogVal(Neff_lm[l][m]) : LogVal(LogVal::LogZero));
lPrecalc_lm[l][m] = lNeff_lm[l][m] * lBinSize * toLogValExp(-controlParameterValues[l] * U_m[m]); //for updateDensityOfStates
Neff_l[l] += Neff_lm[l][m];
}
}
}
inline void MultireweightHistosPT::updateDensityOfStates() {
//precalculated the part that does not depend on the estimates of the partition functions
//lPrecalc_lm[l][m] = lNeff_lm[l][m] * lBinSize * toLogValExp(-betas[l] * U_m[m]);