Subversion Repositories QNX 8.QNX8 LLVM/Clang compiler suite

Rev

Details | Last modification | View Log | RSS feed

Rev Author Line No. Line
14 pmbaty 1
//===- llvm/ADT/SparseMultiSet.h - Sparse multiset --------------*- C++ -*-===//
2
//
3
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4
// See https://llvm.org/LICENSE.txt for license information.
5
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6
//
7
//===----------------------------------------------------------------------===//
8
///
9
/// \file
10
/// This file defines the SparseMultiSet class, which adds multiset behavior to
11
/// the SparseSet.
12
///
13
/// A sparse multiset holds a small number of objects identified by integer keys
14
/// from a moderately sized universe. The sparse multiset uses more memory than
15
/// other containers in order to provide faster operations. Any key can map to
16
/// multiple values. A SparseMultiSetNode class is provided, which serves as a
17
/// convenient base class for the contents of a SparseMultiSet.
18
///
19
//===----------------------------------------------------------------------===//
20
 
21
#ifndef LLVM_ADT_SPARSEMULTISET_H
22
#define LLVM_ADT_SPARSEMULTISET_H
23
 
24
#include "llvm/ADT/identity.h"
25
#include "llvm/ADT/SmallVector.h"
26
#include "llvm/ADT/SparseSet.h"
27
#include <cassert>
28
#include <cstdint>
29
#include <cstdlib>
30
#include <iterator>
31
#include <limits>
32
#include <utility>
33
 
34
namespace llvm {
35
 
36
/// Fast multiset implementation for objects that can be identified by small
37
/// unsigned keys.
38
///
39
/// SparseMultiSet allocates memory proportional to the size of the key
40
/// universe, so it is not recommended for building composite data structures.
41
/// It is useful for algorithms that require a single set with fast operations.
42
///
43
/// Compared to DenseSet and DenseMap, SparseMultiSet provides constant-time
44
/// fast clear() as fast as a vector.  The find(), insert(), and erase()
45
/// operations are all constant time, and typically faster than a hash table.
46
/// The iteration order doesn't depend on numerical key values, it only depends
47
/// on the order of insert() and erase() operations.  Iteration order is the
48
/// insertion order. Iteration is only provided over elements of equivalent
49
/// keys, but iterators are bidirectional.
50
///
51
/// Compared to BitVector, SparseMultiSet<unsigned> uses 8x-40x more memory, but
52
/// offers constant-time clear() and size() operations as well as fast iteration
53
/// independent on the size of the universe.
54
///
55
/// SparseMultiSet contains a dense vector holding all the objects and a sparse
56
/// array holding indexes into the dense vector.  Most of the memory is used by
57
/// the sparse array which is the size of the key universe. The SparseT template
58
/// parameter provides a space/speed tradeoff for sets holding many elements.
59
///
60
/// When SparseT is uint32_t, find() only touches up to 3 cache lines, but the
61
/// sparse array uses 4 x Universe bytes.
62
///
63
/// When SparseT is uint8_t (the default), find() touches up to 3+[N/256] cache
64
/// lines, but the sparse array is 4x smaller.  N is the number of elements in
65
/// the set.
66
///
67
/// For sets that may grow to thousands of elements, SparseT should be set to
68
/// uint16_t or uint32_t.
69
///
70
/// Multiset behavior is provided by providing doubly linked lists for values
71
/// that are inlined in the dense vector. SparseMultiSet is a good choice when
72
/// one desires a growable number of entries per key, as it will retain the
73
/// SparseSet algorithmic properties despite being growable. Thus, it is often a
74
/// better choice than a SparseSet of growable containers or a vector of
75
/// vectors. SparseMultiSet also keeps iterators valid after erasure (provided
76
/// the iterators don't point to the element erased), allowing for more
77
/// intuitive and fast removal.
78
///
79
/// @tparam ValueT      The type of objects in the set.
80
/// @tparam KeyFunctorT A functor that computes an unsigned index from KeyT.
81
/// @tparam SparseT     An unsigned integer type. See above.
82
///
83
template<typename ValueT,
84
         typename KeyFunctorT = identity<unsigned>,
85
         typename SparseT = uint8_t>
86
class SparseMultiSet {
87
  static_assert(std::is_unsigned_v<SparseT>,
88
                "SparseT must be an unsigned integer type");
89
 
90
  /// The actual data that's stored, as a doubly-linked list implemented via
91
  /// indices into the DenseVector.  The doubly linked list is implemented
92
  /// circular in Prev indices, and INVALID-terminated in Next indices. This
93
  /// provides efficient access to list tails. These nodes can also be
94
  /// tombstones, in which case they are actually nodes in a single-linked
95
  /// freelist of recyclable slots.
96
  struct SMSNode {
97
    static constexpr unsigned INVALID = ~0U;
98
 
99
    ValueT Data;
100
    unsigned Prev;
101
    unsigned Next;
102
 
103
    SMSNode(ValueT D, unsigned P, unsigned N) : Data(D), Prev(P), Next(N) {}
104
 
105
    /// List tails have invalid Nexts.
106
    bool isTail() const {
107
      return Next == INVALID;
108
    }
109
 
110
    /// Whether this node is a tombstone node, and thus is in our freelist.
111
    bool isTombstone() const {
112
      return Prev == INVALID;
113
    }
114
 
115
    /// Since the list is circular in Prev, all non-tombstone nodes have a valid
116
    /// Prev.
117
    bool isValid() const { return Prev != INVALID; }
118
  };
119
 
120
  using KeyT = typename KeyFunctorT::argument_type;
121
  using DenseT = SmallVector<SMSNode, 8>;
122
  DenseT Dense;
123
  SparseT *Sparse = nullptr;
124
  unsigned Universe = 0;
125
  KeyFunctorT KeyIndexOf;
126
  SparseSetValFunctor<KeyT, ValueT, KeyFunctorT> ValIndexOf;
127
 
128
  /// We have a built-in recycler for reusing tombstone slots. This recycler
129
  /// puts a singly-linked free list into tombstone slots, allowing us quick
130
  /// erasure, iterator preservation, and dense size.
131
  unsigned FreelistIdx = SMSNode::INVALID;
132
  unsigned NumFree = 0;
133
 
134
  unsigned sparseIndex(const ValueT &Val) const {
135
    assert(ValIndexOf(Val) < Universe &&
136
           "Invalid key in set. Did object mutate?");
137
    return ValIndexOf(Val);
138
  }
139
  unsigned sparseIndex(const SMSNode &N) const { return sparseIndex(N.Data); }
140
 
141
  /// Whether the given entry is the head of the list. List heads's previous
142
  /// pointers are to the tail of the list, allowing for efficient access to the
143
  /// list tail. D must be a valid entry node.
144
  bool isHead(const SMSNode &D) const {
145
    assert(D.isValid() && "Invalid node for head");
146
    return Dense[D.Prev].isTail();
147
  }
148
 
149
  /// Whether the given entry is a singleton entry, i.e. the only entry with
150
  /// that key.
151
  bool isSingleton(const SMSNode &N) const {
152
    assert(N.isValid() && "Invalid node for singleton");
153
    // Is N its own predecessor?
154
    return &Dense[N.Prev] == &N;
155
  }
156
 
157
  /// Add in the given SMSNode. Uses a free entry in our freelist if
158
  /// available. Returns the index of the added node.
159
  unsigned addValue(const ValueT& V, unsigned Prev, unsigned Next) {
160
    if (NumFree == 0) {
161
      Dense.push_back(SMSNode(V, Prev, Next));
162
      return Dense.size() - 1;
163
    }
164
 
165
    // Peel off a free slot
166
    unsigned Idx = FreelistIdx;
167
    unsigned NextFree = Dense[Idx].Next;
168
    assert(Dense[Idx].isTombstone() && "Non-tombstone free?");
169
 
170
    Dense[Idx] = SMSNode(V, Prev, Next);
171
    FreelistIdx = NextFree;
172
    --NumFree;
173
    return Idx;
174
  }
175
 
176
  /// Make the current index a new tombstone. Pushes it onto the freelist.
177
  void makeTombstone(unsigned Idx) {
178
    Dense[Idx].Prev = SMSNode::INVALID;
179
    Dense[Idx].Next = FreelistIdx;
180
    FreelistIdx = Idx;
181
    ++NumFree;
182
  }
183
 
184
public:
185
  using value_type = ValueT;
186
  using reference = ValueT &;
187
  using const_reference = const ValueT &;
188
  using pointer = ValueT *;
189
  using const_pointer = const ValueT *;
190
  using size_type = unsigned;
191
 
192
  SparseMultiSet() = default;
193
  SparseMultiSet(const SparseMultiSet &) = delete;
194
  SparseMultiSet &operator=(const SparseMultiSet &) = delete;
195
  ~SparseMultiSet() { free(Sparse); }
196
 
197
  /// Set the universe size which determines the largest key the set can hold.
198
  /// The universe must be sized before any elements can be added.
199
  ///
200
  /// @param U Universe size. All object keys must be less than U.
201
  ///
202
  void setUniverse(unsigned U) {
203
    // It's not hard to resize the universe on a non-empty set, but it doesn't
204
    // seem like a likely use case, so we can add that code when we need it.
205
    assert(empty() && "Can only resize universe on an empty map");
206
    // Hysteresis prevents needless reallocations.
207
    if (U >= Universe/4 && U <= Universe)
208
      return;
209
    free(Sparse);
210
    // The Sparse array doesn't actually need to be initialized, so malloc
211
    // would be enough here, but that will cause tools like valgrind to
212
    // complain about branching on uninitialized data.
213
    Sparse = static_cast<SparseT*>(safe_calloc(U, sizeof(SparseT)));
214
    Universe = U;
215
  }
216
 
217
  /// Our iterators are iterators over the collection of objects that share a
218
  /// key.
219
  template <typename SMSPtrTy> class iterator_base {
220
    friend class SparseMultiSet;
221
 
222
  public:
223
    using iterator_category = std::bidirectional_iterator_tag;
224
    using value_type = ValueT;
225
    using difference_type = std::ptrdiff_t;
226
    using pointer = value_type *;
227
    using reference = value_type &;
228
 
229
  private:
230
    SMSPtrTy SMS;
231
    unsigned Idx;
232
    unsigned SparseIdx;
233
 
234
    iterator_base(SMSPtrTy P, unsigned I, unsigned SI)
235
      : SMS(P), Idx(I), SparseIdx(SI) {}
236
 
237
    /// Whether our iterator has fallen outside our dense vector.
238
    bool isEnd() const {
239
      if (Idx == SMSNode::INVALID)
240
        return true;
241
 
242
      assert(Idx < SMS->Dense.size() && "Out of range, non-INVALID Idx?");
243
      return false;
244
    }
245
 
246
    /// Whether our iterator is properly keyed, i.e. the SparseIdx is valid
247
    bool isKeyed() const { return SparseIdx < SMS->Universe; }
248
 
249
    unsigned Prev() const { return SMS->Dense[Idx].Prev; }
250
    unsigned Next() const { return SMS->Dense[Idx].Next; }
251
 
252
    void setPrev(unsigned P) { SMS->Dense[Idx].Prev = P; }
253
    void setNext(unsigned N) { SMS->Dense[Idx].Next = N; }
254
 
255
  public:
256
    reference operator*() const {
257
      assert(isKeyed() && SMS->sparseIndex(SMS->Dense[Idx].Data) == SparseIdx &&
258
             "Dereferencing iterator of invalid key or index");
259
 
260
      return SMS->Dense[Idx].Data;
261
    }
262
    pointer operator->() const { return &operator*(); }
263
 
264
    /// Comparison operators
265
    bool operator==(const iterator_base &RHS) const {
266
      // end compares equal
267
      if (SMS == RHS.SMS && Idx == RHS.Idx) {
268
        assert((isEnd() || SparseIdx == RHS.SparseIdx) &&
269
               "Same dense entry, but different keys?");
270
        return true;
271
      }
272
 
273
      return false;
274
    }
275
 
276
    bool operator!=(const iterator_base &RHS) const {
277
      return !operator==(RHS);
278
    }
279
 
280
    /// Increment and decrement operators
281
    iterator_base &operator--() { // predecrement - Back up
282
      assert(isKeyed() && "Decrementing an invalid iterator");
283
      assert((isEnd() || !SMS->isHead(SMS->Dense[Idx])) &&
284
             "Decrementing head of list");
285
 
286
      // If we're at the end, then issue a new find()
287
      if (isEnd())
288
        Idx = SMS->findIndex(SparseIdx).Prev();
289
      else
290
        Idx = Prev();
291
 
292
      return *this;
293
    }
294
    iterator_base &operator++() { // preincrement - Advance
295
      assert(!isEnd() && isKeyed() && "Incrementing an invalid/end iterator");
296
      Idx = Next();
297
      return *this;
298
    }
299
    iterator_base operator--(int) { // postdecrement
300
      iterator_base I(*this);
301
      --*this;
302
      return I;
303
    }
304
    iterator_base operator++(int) { // postincrement
305
      iterator_base I(*this);
306
      ++*this;
307
      return I;
308
    }
309
  };
310
 
311
  using iterator = iterator_base<SparseMultiSet *>;
312
  using const_iterator = iterator_base<const SparseMultiSet *>;
313
 
314
  // Convenience types
315
  using RangePair = std::pair<iterator, iterator>;
316
 
317
  /// Returns an iterator past this container. Note that such an iterator cannot
318
  /// be decremented, but will compare equal to other end iterators.
319
  iterator end() { return iterator(this, SMSNode::INVALID, SMSNode::INVALID); }
320
  const_iterator end() const {
321
    return const_iterator(this, SMSNode::INVALID, SMSNode::INVALID);
322
  }
323
 
324
  /// Returns true if the set is empty.
325
  ///
326
  /// This is not the same as BitVector::empty().
327
  ///
328
  bool empty() const { return size() == 0; }
329
 
330
  /// Returns the number of elements in the set.
331
  ///
332
  /// This is not the same as BitVector::size() which returns the size of the
333
  /// universe.
334
  ///
335
  size_type size() const {
336
    assert(NumFree <= Dense.size() && "Out-of-bounds free entries");
337
    return Dense.size() - NumFree;
338
  }
339
 
340
  /// Clears the set.  This is a very fast constant time operation.
341
  ///
342
  void clear() {
343
    // Sparse does not need to be cleared, see find().
344
    Dense.clear();
345
    NumFree = 0;
346
    FreelistIdx = SMSNode::INVALID;
347
  }
348
 
349
  /// Find an element by its index.
350
  ///
351
  /// @param   Idx A valid index to find.
352
  /// @returns An iterator to the element identified by key, or end().
353
  ///
354
  iterator findIndex(unsigned Idx) {
355
    assert(Idx < Universe && "Key out of range");
356
    const unsigned Stride = std::numeric_limits<SparseT>::max() + 1u;
357
    for (unsigned i = Sparse[Idx], e = Dense.size(); i < e; i += Stride) {
358
      const unsigned FoundIdx = sparseIndex(Dense[i]);
359
      // Check that we're pointing at the correct entry and that it is the head
360
      // of a valid list.
361
      if (Idx == FoundIdx && Dense[i].isValid() && isHead(Dense[i]))
362
        return iterator(this, i, Idx);
363
      // Stride is 0 when SparseT >= unsigned.  We don't need to loop.
364
      if (!Stride)
365
        break;
366
    }
367
    return end();
368
  }
369
 
370
  /// Find an element by its key.
371
  ///
372
  /// @param   Key A valid key to find.
373
  /// @returns An iterator to the element identified by key, or end().
374
  ///
375
  iterator find(const KeyT &Key) {
376
    return findIndex(KeyIndexOf(Key));
377
  }
378
 
379
  const_iterator find(const KeyT &Key) const {
380
    iterator I = const_cast<SparseMultiSet*>(this)->findIndex(KeyIndexOf(Key));
381
    return const_iterator(I.SMS, I.Idx, KeyIndexOf(Key));
382
  }
383
 
384
  /// Returns the number of elements identified by Key. This will be linear in
385
  /// the number of elements of that key.
386
  size_type count(const KeyT &Key) const {
387
    unsigned Ret = 0;
388
    for (const_iterator It = find(Key); It != end(); ++It)
389
      ++Ret;
390
 
391
    return Ret;
392
  }
393
 
394
  /// Returns true if this set contains an element identified by Key.
395
  bool contains(const KeyT &Key) const {
396
    return find(Key) != end();
397
  }
398
 
399
  /// Return the head and tail of the subset's list, otherwise returns end().
400
  iterator getHead(const KeyT &Key) { return find(Key); }
401
  iterator getTail(const KeyT &Key) {
402
    iterator I = find(Key);
403
    if (I != end())
404
      I = iterator(this, I.Prev(), KeyIndexOf(Key));
405
    return I;
406
  }
407
 
408
  /// The bounds of the range of items sharing Key K. First member is the head
409
  /// of the list, and the second member is a decrementable end iterator for
410
  /// that key.
411
  RangePair equal_range(const KeyT &K) {
412
    iterator B = find(K);
413
    iterator E = iterator(this, SMSNode::INVALID, B.SparseIdx);
414
    return std::make_pair(B, E);
415
  }
416
 
417
  /// Insert a new element at the tail of the subset list. Returns an iterator
418
  /// to the newly added entry.
419
  iterator insert(const ValueT &Val) {
420
    unsigned Idx = sparseIndex(Val);
421
    iterator I = findIndex(Idx);
422
 
423
    unsigned NodeIdx = addValue(Val, SMSNode::INVALID, SMSNode::INVALID);
424
 
425
    if (I == end()) {
426
      // Make a singleton list
427
      Sparse[Idx] = NodeIdx;
428
      Dense[NodeIdx].Prev = NodeIdx;
429
      return iterator(this, NodeIdx, Idx);
430
    }
431
 
432
    // Stick it at the end.
433
    unsigned HeadIdx = I.Idx;
434
    unsigned TailIdx = I.Prev();
435
    Dense[TailIdx].Next = NodeIdx;
436
    Dense[HeadIdx].Prev = NodeIdx;
437
    Dense[NodeIdx].Prev = TailIdx;
438
 
439
    return iterator(this, NodeIdx, Idx);
440
  }
441
 
442
  /// Erases an existing element identified by a valid iterator.
443
  ///
444
  /// This invalidates iterators pointing at the same entry, but erase() returns
445
  /// an iterator pointing to the next element in the subset's list. This makes
446
  /// it possible to erase selected elements while iterating over the subset:
447
  ///
448
  ///   tie(I, E) = Set.equal_range(Key);
449
  ///   while (I != E)
450
  ///     if (test(*I))
451
  ///       I = Set.erase(I);
452
  ///     else
453
  ///       ++I;
454
  ///
455
  /// Note that if the last element in the subset list is erased, this will
456
  /// return an end iterator which can be decremented to get the new tail (if it
457
  /// exists):
458
  ///
459
  ///  tie(B, I) = Set.equal_range(Key);
460
  ///  for (bool isBegin = B == I; !isBegin; /* empty */) {
461
  ///    isBegin = (--I) == B;
462
  ///    if (test(I))
463
  ///      break;
464
  ///    I = erase(I);
465
  ///  }
466
  iterator erase(iterator I) {
467
    assert(I.isKeyed() && !I.isEnd() && !Dense[I.Idx].isTombstone() &&
468
           "erasing invalid/end/tombstone iterator");
469
 
470
    // First, unlink the node from its list. Then swap the node out with the
471
    // dense vector's last entry
472
    iterator NextI = unlink(Dense[I.Idx]);
473
 
474
    // Put in a tombstone.
475
    makeTombstone(I.Idx);
476
 
477
    return NextI;
478
  }
479
 
480
  /// Erase all elements with the given key. This invalidates all
481
  /// iterators of that key.
482
  void eraseAll(const KeyT &K) {
483
    for (iterator I = find(K); I != end(); /* empty */)
484
      I = erase(I);
485
  }
486
 
487
private:
488
  /// Unlink the node from its list. Returns the next node in the list.
489
  iterator unlink(const SMSNode &N) {
490
    if (isSingleton(N)) {
491
      // Singleton is already unlinked
492
      assert(N.Next == SMSNode::INVALID && "Singleton has next?");
493
      return iterator(this, SMSNode::INVALID, ValIndexOf(N.Data));
494
    }
495
 
496
    if (isHead(N)) {
497
      // If we're the head, then update the sparse array and our next.
498
      Sparse[sparseIndex(N)] = N.Next;
499
      Dense[N.Next].Prev = N.Prev;
500
      return iterator(this, N.Next, ValIndexOf(N.Data));
501
    }
502
 
503
    if (N.isTail()) {
504
      // If we're the tail, then update our head and our previous.
505
      findIndex(sparseIndex(N)).setPrev(N.Prev);
506
      Dense[N.Prev].Next = N.Next;
507
 
508
      // Give back an end iterator that can be decremented
509
      iterator I(this, N.Prev, ValIndexOf(N.Data));
510
      return ++I;
511
    }
512
 
513
    // Otherwise, just drop us
514
    Dense[N.Next].Prev = N.Prev;
515
    Dense[N.Prev].Next = N.Next;
516
    return iterator(this, N.Next, ValIndexOf(N.Data));
517
  }
518
};
519
 
520
} // end namespace llvm
521
 
522
#endif // LLVM_ADT_SPARSEMULTISET_H