File: secondarycohom.cpp

package info (click to toggle)
cohomcalg 0.32%2Bds-6
  • links: PTS, VCS
  • area: main
  • in suites: bookworm, sid, trixie
  • size: 2,008 kB
  • sloc: cpp: 3,291; makefile: 46; ansic: 17
file content (796 lines) | stat: -rw-r--r-- 30,958 bytes parent folder | download | duplicates (3)
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
////////////////////////////////////////////////////////////////////////////////////////////////////
//                                                                                                //
//  secondarycohom.cpp                                                                            //
//  ==================                                                                            //
//                                                                                                //
//  Code: Benjamin Jurke, http://benjaminjurke.net                                                //
//                                                                                                //
////////////////////////////////////////////////////////////////////////////////////////////////////
//                                                                                                //
//  File history:                                                                                 //
//        - 04.04.2010  File created as secondarycohom.cpp                                        //
//                      Handles the computations of the secondary cohomology which ultimately     //
//                      yields a list of the relevant denominator monomials and their respective  //
//                      multiplicities.                                                           //
//                                                                                                //
////////////////////////////////////////////////////////////////////////////////////////////////////


#include <iterator>
#include <fstream>
#include <ctime>
#include <cstring>
#include <thread>

#include "secondarycohom.h"
#include "main.h"
#include "platform.h"


using namespace std;


////////////////////////////////////////////////////////////////////////////////////////////////////


#define TRAVERSE_SMP_THRESHOLD   8192
#define INSERTION_BUFFERSIZE     (1024*1024)      // Size of the insertion buffer window length
#define MONOMIAL_MAP_MAX_FULL_OUTPUT 250      // Max. number of monomials printed with shorten_output


////////////////////////////////////////////////////////////////////////////////////////////////////


void CSecondaryCohomology::Clear()
{
    insertion_prototype.Clear(0);
    unique_monoms.clear();
    min_c_deg = 0;
    max_c_deg = 0;
    len_c_deg = 0;
    monoms.Clear();
}

CSecondaryCohomology::CSecondaryCohomology()
{
    Clear();
}

string CSecondaryCohomology::Print2ndSeq(const MonomData &mondat, int32_t min_out, int32_t max_out)
{
    /* This output function prints the C-degrees of a secondary/remnant sequence in the
       range C^min_out to C^max_out. The output range values are checked for validity.
       Zero C-degrees in this list are replaced by spaces to make tabular output possible. */

    string strTmp;
    char buf[64];

    // Check the C-degree range
    if (min_out < min_c_deg)
        min_out = min_c_deg;
    if (max_out > max_c_deg)
        max_out = max_c_deg;

    // Loop through the range and print any non-zero C-degree, otherwise spaces
    for (int32_t j=min_out; j<=max_out; j++)
    {
        safe_sprintf(buf, sizeof(buf), "C^%d=%lld  ", (int) j, (long long int) mondat.c[j-min_c_deg]);
        if (mondat.c[j-min_c_deg] != 0)
            strTmp += buf;
        else
            strTmp += string(strlen(buf), ' ');
    }
    return strTmp;
}

string CSecondaryCohomology::Print2ndSeq(const MonomData &mondat)
{
    /* This output functions prints the C-degrees of the secondary/remnant sequence starting
       with the first and ending with the last non-zero C-degree. Note that this is
       NOT suitable for tabular output. */

    size_t seq_start, seq_end;
    if (GetIntSequenceStartEnd(mondat.c, seq_start, seq_end))
        return Print2ndSeq(mondat, (int32_t) seq_start+min_c_deg, (int32_t) seq_end+min_c_deg);
    else
        return "zero sequence";
}

string CSecondaryCohomology::PrintMonomialWith2ndSeq(const CInternalData &id, MonomMap::const_iterator monom, int32_t min_out, int32_t max_out)
{
    /* This output function prints the monomial (and potentially the internal bit-mask), determines
       the length of the non-zero part of the C-degrees and prints the range of C-degrees from
       min_out to max_out. */

    char buf[64];
    size_t seq_start = 0, seq_end = 0;
    safe_sprintf(buf, sizeof(buf), " with %4ld total (len %2d):   ", (long int) monom->second.num, (int) (seq_end - seq_start + 1));

    // Check if we have a zero sequence
    if (GetIntSequenceStartEnd(monom->second.c, seq_start, seq_end))
        return id.Int64ToMonomialPadded(monom->first) + buf + Print2ndSeq(monom->second, min_out, max_out);
    else
        return id.Int64ToMonomialPadded(monom->first) + buf + "zero sequence";
}


void CSecondaryCohomology::PrintMonomMap(const CInternalData &id, bool shorten_output)
{
    /* This output function prints the entire monomial map in tabular form. Only the necessary
       part of the secondary/remnant sequences is printed. Furthermore an automatic shortening
       of excessive large outputs can be activated. */

    // Determine the smallest and highest non-trivial c-degree in the entire list
    int32_t min_deg = len_c_deg;
    int32_t max_deg = 0;
    for (MonomMap::const_iterator set_iter = unique_monoms.begin(); set_iter != unique_monoms.end(); set_iter++)
    {
        int32_t i;

        // Check FORWARDS from the beginning
        for (i = 0; i<min_deg; i++)
        {
            if (set_iter->second.c[i] != 0)
            {
                min_deg = i;
                break;
            }
        }

        // Check BACKWARDS from the end
        for (i = len_c_deg-1; i >= max_deg; i--)
        {
            if (set_iter->second.c[i] != 0)
            {
                max_deg = i+1;
                break;
            }
        }
    }
    min_deg += min_c_deg;
    max_deg += min_c_deg;

    // Now print the monomials
    size_t numelems = unique_monoms.size();
    MonomMap::const_iterator set_iter = unique_monoms.begin();
    MSG_OUT("There are " << numelems << " elements in the monomial map (using " << (len_c_deg * sizeof(int64_t) * numelems + sizeof(int64_t) + sizeof(uint64_t)) << " bytes of memory):");
    for (size_t i=0; i<numelems; i++)
    {
        MSG_OUT("    " << PrintMonomialWith2ndSeq(id, set_iter, min_deg, max_deg));
        set_iter++;

        // Check if the output should be shortened
        if (shorten_output)
        {
            if (numelems > MONOMIAL_MAP_MAX_FULL_OUTPUT)
            {
                if ((i >= 50) && (i < numelems - 50))
                {
                    MSG_OUT("(.....)");
                    i = numelems - 50;
                }
            }
        }
    }
}


bool CSecondaryCohomology::IsSequenceExact(const vector<uint64_t> &seq, size_t &pos_out) const
{
    /* This computational helper function checks a arbitrary length sequence for naive
       exactness. It also computes the naive kernel via the alternating sum method from
       the end and put out a potential repair position in case of non-exactness. */

    size_t seqlen = seq.size();
    if (seqlen > 0)
    {
        int64_t kernel = seq[seqlen-1];
        for (size_t i=0; i<seqlen; i++)
        {
            kernel = seq[seqlen-(i+1)] - kernel;
            if (kernel < 0)
            {
                pos_out = seqlen - (i+1)+1 + min_c_deg;
                return false;
            }
        }
        return true;
    }
    else
    {
        // A zero-length sequence - perhaps throw some runtime error?!
        WARN_OUT("Internal error while checking secondary sequence exactness.");
        return true;
    }
}


/////////////////////////
// The big main functions


bool CSecondaryCohomology::Init2ndSequences(const CInternalData &id)
{
    /* This function initializes the CSecondaryStructure and computes the theoretical
       minimal and maximal C-degree level. It also prepares the corresponding insertion
       prototype for speedup. */

    // Clear the list of unique monomials resulting from unions;
    unique_monoms.clear();

    // Determine the min/max values of the secondary complex
    min_c_deg = 1 - (int32_t) id.GetNumSRgens();
    max_c_deg = (int32_t) id.GetNumCoordinates();
    len_c_deg = max_c_deg - min_c_deg + 1;

    // Prepare the insertion prototype dummy for speedup
    insertion_prototype.Clear(len_c_deg);

    return true;
}


struct FastMonomialData
{
    uint64_t monomial;
    int nka;

  public:
    friend inline bool operator<(const FastMonomialData &lhs, const FastMonomialData &rhs) { return lhs.monomial < rhs.monomial; };
};

void SortAndFlushBuffer(vector<FastMonomialData> &buffer, MonomMap &unique_monoms, size_t max_run, MonomData &insertion_prototype, int min_c_deg)
{
    /* This internal helper function flushes the monomial insertion buffer. As the monomials
       are kept in an associate list, it helps to presort the monomials, such that the number
       of look-ups in the associative index is minimized. */

    if (max_run > 0)
    {
        // Sort the insertion buffer
        sort(buffer.begin(), buffer.begin()+max_run);

        // The first one we have to get by hand
        pair<MonomMap::iterator, bool> ret = unique_monoms.insert(pair<uint64_t, MonomData>(buffer[0].monomial, insertion_prototype));
        ret.first->second.num++;
        ret.first->second.c[buffer[0].nka - min_c_deg]++;

        // Now loop through the buffer
        for (size_t i=1; i<max_run; i++)
        {
            // Check if the monomial is still the same, otherwise get a new one
            if (ret.first->first != buffer[i].monomial)
                ret = unique_monoms.insert(pair<uint64_t, MonomData>(buffer[i].monomial, insertion_prototype));

            // Increase data
            ret.first->second.num++;
            ret.first->second.c[buffer[i].nka - min_c_deg]++;
        }
    }
}


struct TraverseWorkerData
{
	unsigned int worker_id;
	MonomMap unique_monoms;
	vector<FastMonomialData> vBuffer;
	uint64_t window_start, window_end, window_width, buffer_runs;
	MonomData *insertion_prototype;
	int32_t min_c_deg;
	vector<uint64_t> SRunion[4];
	vector<double> *Status;
};

void TraverseWorkerFunction(void *p_dat)
{
	TraverseWorkerData *wd = (TraverseWorkerData *) p_dat;
	FastMonomialData *pbuf;

	// This is the expensive main loop
    for (uint64_t buffer_loop = 0; buffer_loop < wd->buffer_runs; buffer_loop++)
    {
		uint64_t window_start = wd->window_start + buffer_loop * INSERTION_BUFFERSIZE;
        uint64_t window_end = window_start + INSERTION_BUFFERSIZE - 1;
		if (window_end > wd->window_end)
			window_end = wd->window_end;

        pbuf = &wd->vBuffer[0];

        for (uint64_t sr_combi=window_start; sr_combi<=window_end; sr_combi++, pbuf++)
        {
            uint64_t genunion = wd->SRunion[0][sr_combi & 0xffffu] | wd->SRunion[1][(sr_combi >> 16) & 0xffffu] | wd->SRunion[2][(sr_combi >> 32) & 0xffffu] | wd->SRunion[3][(sr_combi >> 48) & 0xffffu];
            pbuf->monomial = genunion;
            pbuf->nka = CBits::CountBits(genunion) - CBits::CountBits(sr_combi);
        }

		SortAndFlushBuffer(wd->vBuffer, wd->unique_monoms, (size_t) (window_end-window_start + 1), *wd->insertion_prototype, wd->min_c_deg);

        // Show some progress status output
		(*wd->Status)[wd->worker_id] = (double) (window_end - wd->window_start) / wd->window_width;
    }
}

struct StatusWorkerData {
	vector<double> *status;
	clock_t start;
	bool Run;
};

void StatusOutputWorker(void *p_dat)
{
	StatusWorkerData *wstat = (StatusWorkerData *) p_dat;
	if (!wstat)
		return;

	while (wstat->Run)
	{
		double dTotal = 0.0;
		for (size_t i=0; i<wstat->status->size(); i++)
			dTotal += (*wstat->status)[i];
		dTotal /= wstat->status->size();
		double dSecElapsed = (double) (clock() - wstat->start) / CLOCKS_PER_SEC;
		double dSecRemaining = (dTotal > 0.0) ? ((1.0 - dTotal) / dTotal) * dSecElapsed : 1.0;
		dTotal *= 100;
		
		char buf[64];
		safe_sprintf(buf, sizeof(buf), "%.2f", dTotal);
		CONSOLE_OUT("  " << buf << "% completed (" << SecondsToTime((clock_t) dSecRemaining) << " remaining)...");

		CONSOLE_OUT("            \r");

		SleepMilliSec(100);
	}

	delete wstat;
}


bool CSecondaryCohomology::TraverseSRpowerset(const CInternalData &id)
{
    /* This is the main powerset traverse function, which loops through the entire range
       of subsets of the Stanley-Reisner generators and computes the corresponding
       monomial unions as well as the secondary/remnant sequences. */

    // As an enormous speedup, we are using a union buffer analogous to the counting of the bits
    // Since we are dealing with 64 bit variables, we are using 4 bit buffer for 16-bit blocks, such
    // that the buffer size is 4*8*64 KB = 2 MBr. Note that this buffer has to be allocated
    // dynamically, otherwise the MS VS compiler throws a stack overflow in debug mode.


	// Prepare and setup the worker threads
	unsigned int numWorkers = thread::hardware_concurrency();
	if (numWorkers == 0)
		numWorkers = 4;

    // Compute the maximal SR bitmask value and the number of buffer runs
    size_t number_of_srgens = id.GetNumSRgens();
    uint64_t sr_powerset_max = (0x1ull << number_of_srgens);
	if (sr_powerset_max < TRAVERSE_SMP_THRESHOLD)
		numWorkers = 1;
	uint64_t sr_powerset_max_per_worker = sr_powerset_max / numWorkers;

    // Allocate the insertion buffer (typically around 12-16 MB)
	TraverseWorkerData *workerdat = new TraverseWorkerData[numWorkers];
	vector<uint64_t> paddedSRvec(id.GetSRgens());
	paddedSRvec.resize(64, 0);
	vector<double> WorkerStatus;
	WorkerStatus.resize(numWorkers, 0.0);
	for (unsigned int k = 0; k < numWorkers; k++)
	{
		workerdat[k].vBuffer.resize(INSERTION_BUFFERSIZE);
		workerdat[k].window_start = k*sr_powerset_max_per_worker;
		workerdat[k].worker_id = k;
		workerdat[k].window_end = (k+1)*sr_powerset_max_per_worker - 1;
		if (k==numWorkers-1)
			workerdat[k].window_end = sr_powerset_max - 1;
		workerdat[k].window_width = workerdat[k].window_end - workerdat[k].window_start + 1;
		workerdat[k].buffer_runs = workerdat[k].window_width / INSERTION_BUFFERSIZE;
		if (workerdat[k].window_width % INSERTION_BUFFERSIZE != 0)
			workerdat[k].buffer_runs++;
		workerdat[k].insertion_prototype = &insertion_prototype;
		workerdat[k].min_c_deg = min_c_deg;
		workerdat[k].Status = &WorkerStatus;

		// Init the worker buffer
		for (size_t i=0; i<4; i++)
		{
			workerdat[k].SRunion[i].resize(0x1u << 16);
			for (size_t j=0; j<(0x1u << 16); j++)
			{
				uint64_t combi = j << (i*16);
				uint64_t genunion = 0;
				for (size_t k=0; k<16; k++)
				{
					if (combi & (0x1ull << (k + i*16)))
						genunion |= paddedSRvec[k + i*16];
				}
				workerdat[k].SRunion[i][j] = genunion;
			}
		}
		
	}
	paddedSRvec.clear();

	// Start the workers
	vector<thread *> worker_threads;
	for (size_t k = 0; k<numWorkers; k++)
	{
		thread *t = new thread(TraverseWorkerFunction, (void*) &workerdat[k]);
		worker_threads.push_back(t);		
	}
	
	StatusWorkerData *wstat = new StatusWorkerData;
	if (!wstat)
		return false;
	wstat->Run = true;
	wstat->status = &WorkerStatus;
	wstat->start = clock();
	thread *stat = new thread(StatusOutputWorker, wstat);

	// Wait for the threads to complete...
	for (size_t k = 0; k<numWorkers; k++)
	{
		worker_threads[k]->join();
		delete worker_threads[k];
	}
	wstat->Run = false;
	

	// Piece together the individually generated data
	unique_monoms.clear();
	for (size_t k = 0; k<numWorkers; k++)
	{
		MonomMap::iterator it = workerdat[k].unique_monoms.begin();

		while (it != workerdat[k].unique_monoms.end())
		{
			pair<MonomMap::iterator, bool> ret = unique_monoms.insert(pair<uint64_t, MonomData>(it->first, it->second));
			if (ret.second==false)
			{
				size_t chain_len = it->second.c.size();
				ret.first->second.num += it->second.num;
				for (size_t r = 0; r < chain_len; r++)
					ret.first->second.c[r] += it->second.c[r];
			}
			it++;
		}
	}

    CONSOLE_OUT(string(75, ' ') << "\r");

    return true;
}


bool CSecondaryCohomology::Compute2ndCohomFromTrivialSequences()
{
    /* After generating the secondary/remnant sequences, the secondary/remnant "cohomology"
       has to be computed from this data. However, lacking a proper description of the mappings
       one has to take a couple of assumptions and employ the Serre-duality to resolve a couple
       of ambiguities. For now, we are taking a two-step scan through the entire range of the
       computed monomials, where in the first one we are simply removing the trivial cases:
       If only a single number is not equal to 0, the sequence is non-exact, but the remnant
       cohomology is obvious. Furthermore, we remove all obviously exact sequence, which is
       quickly checked by computing the alternating sum. */

    MonomMap::iterator cur = unique_monoms.begin();
    while (cur != unique_monoms.end())
    {
        if (cur->second.num == 1)
        {
            // Take care of a sequence containing just a single value 1
            for (size_t i=0; i<len_c_deg; i++)
            {
                if (cur->second.c[i] != 0)
                {
                    if (!monoms.AddUniqueContribution(cur->first, (int32_t) i+min_c_deg, 1))
                        return false;
                    break;
                }
            }
            unique_monoms.erase(cur++);
        }
        else
        {
            // Compute the alternating sum of the secondary/remnant sequence
            int64_t altersum = 0;
            for (size_t i=0; i<len_c_deg; i++)
            {
                if (i % 2)
                    altersum += cur->second.c[i];
                else
                    altersum -= cur->second.c[i];
            }

            if (altersum == 0)
            {
                // Sequence is exact, so discard
                unique_monoms.erase(cur++);
            }
            else
            {
                // Sequence is non-exact & non-trivial, so keep the altersum value AND compute the sequence length
                uint32_t seq_start = len_c_deg;
                for (uint32_t i=0; i<len_c_deg; i++)
                {
                    if (cur->second.c[i] != 0)
                    {
                        seq_start = i;
                        break;
                    }
                }

                uint32_t seq_end = 0;
                for (uint32_t i=len_c_deg-1; i>=0; i--)
                {
                    if (cur->second.c[i] != 0)
                    {
                        seq_end = i;
                        break;
                    }
                }

                uint32_t seq_len = seq_end - seq_start + 1;
                if (seq_len < 1)
                {
                    // This would imply a complete zero sequence!
                    ERR_OUT("Internal error while computing secondary cohomology.");
                    return false;
                }

                // Take care of special length
                switch (seq_len)
                {
                    case 1:
                        // Sequence has a single non-zero value
                        if (!monoms.AddUniqueContribution(cur->first, seq_start+min_c_deg, (uint32_t) cur->second.c[seq_start]))
                            return false;

                        break;

                    default:
                        // Otherwise, store the alternating sum value
                        cur->second.altersum = altersum;
                        cur++;
                }
            }
        }
    }

    return true;
}


bool CSecondaryCohomology::Compute2ndCohomFromSequences(const CInternalData &id)
{
    /* In the second part of the secondary/remnant cohomology computation, we are taking care
       of the non-trivial cases remaining. In principle this procedure is suitable for all
       cases, however, it is somewhat more expensive, therefore the first run to remove the
       "obvious" secondary/remnant sequences. The basic idea is to "repair" the secondary
       sequence in the C-degree range of 0...dim in order to make it exact again. The repair
       value and its possible positions then yield the cohomology of this sequence. On the
       other hand, this is ultimately the origin of the troublesome ambiguities, which can
       only be properly removed by understanding the mappings between the C-degrees. */

    // The temporary "repair" storage
    vector<CohomContrib> fixes;

    // Loop through all remaining monomials in the list
    size_t variety_dim = id.GetDimension();
    MonomMap::iterator cur = unique_monoms.begin();
    while (cur != unique_monoms.end())
    {
        if (cur->second.altersum != 0)
        {
            uint64_t absaltersum = abs(cur->second.altersum);
            fixes.clear();

            // Search for the the C-degrees in range, which is >= the absolute value of the
            // alternating sum of the sequence, which measures the "total failure" exactness
            for (unsigned int j=0; j<=variety_dim; j++)
            {
                if (cur->second.c[j-min_c_deg] >= absaltersum)
                {
                    // Now descrease the value of this position by the absolute value of the
                    // alternating sum and check, if this makes the sequence exact
                    vector<uint64_t> tempseq(cur->second.c);
                    tempseq[j-min_c_deg] -= absaltersum;
                    size_t dummy;
                    if (IsSequenceExact(tempseq, dummy))
                    {
                        // If exactness can be restored this way, the cohomology group at the
                        // respective repair position has the dimension of the absolute value
                        // of the alternating sum - we therefore store this fix as a candidate
                        CohomContrib newfix;
                        newfix.nGroup = j;
                        newfix.nFactor = (uint32_t) absaltersum;
                        fixes.push_back(newfix);
                    }
                }
            }

            // Check if we have obtained at least one "repair" value / secondary cohomology group
            size_t num_fixes = fixes.size();
            if (num_fixes < 1)
            {
                // This should NEVER happen, but is potentially possible, as the
                // repair range is restricted to 0...dim instead of the total sequence
                ERR_OUT("FATAL: Could no properly resolve the exactness of a secondary sequence");
                MSG_OUT(Print2ndSeq(cur->second));
                return false;
            }

            // Now, do we have multi-contributions or not?
            if (num_fixes > 1)
            {
                // Ambiguous contribution we have to take care of lateron
                if (!monoms.AddAmbiguousContribution(cur->first, fixes))
                    return false;
            }
            else
            {
                // Unique contribution... that's fine
                if (!monoms.AddUniqueContribution(cur->first, fixes[0].nGroup, fixes[0].nFactor))
                    return false;
            }
        }
        unique_monoms.erase(cur++);
    }

    // At this point there should be no more monomials left, but check anyway
    if (unique_monoms.size() > 0)
    {
        ERR_OUT("Some monomials could not be properly processed.");
        return false;
    }

    // Clear up allocated buffer space of the monomials (should not be too much...)
    unique_monoms.clear();

    return true;
}


bool CSecondaryCohomology::Perform2ndCohomSerreReduction(const CInternalData &id)
{
    /* Now we have to take care of the ambiguos contributions derive from from restoring the
       exactness of the secondary/remnant sequence. The idea is to check if the complement
       monomial can already be found in the list of unique contributions, because this would
       obviously correspond to a Serre-dual contribution. If so, we can turn the ambiguous
       contribution in an unique one. However, this does not solve every case. In the second
       scan, we check if from the remaining ambiguous contributions an ambiguous complement
       monomial can be found, whose ambiguous contributions might be just right to identify
       a unique one. Unfortunately, this still leaves some cases behind, which we have to deal
       with in the actual computation of the cohomology dimensions. */

    size_t num_ambiguous_before = monoms.ambiguous_monoms.size();
    uint64_t complete_union = id.GetCompleteUnion();
    uint32_t variety_dim = (uint32_t) id.GetDimension();

    // Scan through all ambiguous contributions and check if the complement monomial is unique
    for (map<uint64_t, AmbiguousContribData>::iterator ita = monoms.ambiguous_monoms.begin(); ita != monoms.ambiguous_monoms.end(); )
    {
        uint64_t monomial = ita->first;
        uint64_t complement = (~monomial) & complete_union;

        // First try to find a UNIQUE complement partner
        map<uint64_t, UniqueContribData>::const_iterator uc = monoms.unique_monoms.find(complement);
        if (uc != monoms.unique_monoms.end())
        {
            // Check if this complement partner has a compatible contribution, which would identify
            // a correct contribution of our ambiguous case
            uint32_t group_suggestion = variety_dim - uc->second.Cohom.nGroup;
            size_t numcandidates = ita->second.vCohoms.size();
            bool bFoundCandidate = false;
            size_t candidate = 0;
            for (size_t i=0; i<numcandidates; i++)
            {
                if (ita->second.vCohoms[i].nGroup == group_suggestion)
                {
                    bFoundCandidate = true;
                    candidate = i;
                    break;
                }
            }

            if (bFoundCandidate)
            {
                // So the unique complement monomial exists and identifies the correct contribution
                UniqueContribData newunique;
                newunique.Cohom.nGroup = group_suggestion;
                newunique.Cohom.nFactor = ita->second.vCohoms[candidate].nFactor;
                newunique.nRationals = ita->second.nRationals;
                newunique.nRationalsDual = ita->second.nRationalsDual;

                // Now add the identified unique one
                pair<map<uint64_t, UniqueContribData>::iterator, bool> nu = monoms.unique_monoms.insert(pair<uint64_t, UniqueContribData>(monomial, newunique));
                if (nu.second == false)
                {
                    ERR_OUT("The monomial " << id.Int64ToMonomial(monomial) << " already exists in the unique monomials list!");
                    return false;
                }

                // Remove the ambiguous contribution and start the scan again at the beginning
                monoms.ambiguous_monoms.erase(ita);
                ita = monoms.ambiguous_monoms.begin();
            }
            else
                ita++;
        }
        else
        {
            // If no UNIQUE partner can be found, try to find an ambiguous partner...
            map<uint64_t, AmbiguousContribData>::iterator ita2 = monoms.ambiguous_monoms.find(complement);
            if (ita2 != monoms.ambiguous_monoms.end())
            {
                // OK, we have a partner. Then try to find a "Serre-dual" intersection of the potential contributions
                vector<CohomContrib> intersection;
                size_t cohoms_ita  = ita->second.vCohoms.size();
                size_t cohoms_ita2 = ita2->second.vCohoms.size();
                intersection.clear();
                for (size_t i=0; i<cohoms_ita; i++)
                {
                    for (size_t j=0; j<cohoms_ita2; j++)
                    {
                        if (ita->second.vCohoms[i].nGroup == (variety_dim - ita2->second.vCohoms[j].nGroup))
                        {
                            // Store each such intersection of complement contributions
                            intersection.push_back(ita->second.vCohoms[i]);
                            intersection.push_back(ita2->second.vCohoms[j]);
                        }
                    }
                }

                // Check if the intersection was able to uniquely identify a contribution cohomology group
                if (intersection.size() == 2)
                {
                    // Apparently, we could uniquely resolve the issue

                    // Add the "normal" monomial to the list of unique ones
                    UniqueContribData newunique;
                    newunique.Cohom.nGroup = intersection[0].nGroup;
                    newunique.Cohom.nFactor = intersection[0].nFactor;
                    newunique.nRationals = ita->second.nRationals;
                    newunique.nRationalsDual = ita->second.nRationalsDual;
                    pair<map<uint64_t, UniqueContribData>::iterator, bool> ret = monoms.unique_monoms.insert(std::pair<uint64_t, UniqueContribData>(monomial, newunique));

                    // Add the "complement" partner monomial to the list of unique ones
                    newunique.Cohom.nGroup = intersection[1].nGroup;
                    newunique.Cohom.nFactor = intersection[1].nFactor;
                    newunique.nRationals = ita2->second.nRationals;
                    newunique.nRationalsDual = ita2->second.nRationalsDual;
                    pair<map<uint64_t, UniqueContribData>::iterator, bool> ret2 = monoms.unique_monoms.insert(pair<uint64_t, UniqueContribData>(complement, newunique));

                    if ((ret.second == false) || (ret2.second == false))
                    {
                        ERR_OUT("INTERNAL: Could not add resolved ambiguous monomials to unique monomials list.");
                        return false;
                    }

                    // Delete both ambiguous cases from the ambiguous list & start the scan from the beginning
                    monoms.ambiguous_monoms.erase(monomial);
                    monoms.ambiguous_monoms.erase(complement);
                    ita = monoms.ambiguous_monoms.begin();
                }
                else
                    ita++;
            }
			else
			{
				// CONSOLE_MSG_OUT("WARNING: The Serre duality check could not entirely resolve all monomials.");
				ita++;
			}
        }
    }

    // In the best case scenario, we could resolve all ambiguous cases and turn them into
    // uniques. Unfortunately, this is not always the case.
    size_t num_ambiguous_after = monoms.ambiguous_monoms.size();
    if (num_ambiguous_after > 0)
    {
        WARN_OUT("The Serre dualization reduction was unable to uniquely resolve " << num_ambiguous_after << " of the original " << num_ambiguous_before << " ambiguous monoms.");
    }

    return true;
}