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
|
/* SparseSet implementation.
Copyright (C) 2007-2023 Free Software Foundation, Inc.
Contributed by Peter Bergner <bergner@vnet.ibm.com>
This file is part of GCC.
GCC is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 3, or (at your option) any later
version.
GCC is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License
along with GCC; see the file COPYING3. If not see
<http://www.gnu.org/licenses/>. */
#ifndef GCC_SPARSESET_H
#define GCC_SPARSESET_H
/* Implementation of the Briggs and Torczon sparse set representation.
The sparse set representation was first published in:
"An Efficient Representation for Sparse Sets",
ACM LOPLAS, Vol. 2, Nos. 1-4, March-December 1993, Pages 59-69.
The sparse set representation is suitable for integer sets with a
fixed-size universe. Two vectors are used to store the members of
the set. If an element I is in the set, then sparse[I] is the
index of I in the dense vector, and dense[sparse[I]] == I. The dense
vector works like a stack. The size of the stack is the cardinality
of the set.
The following operations can be performed in O(1) time:
* clear : sparseset_clear
* cardinality : sparseset_cardinality
* set_size : sparseset_size
* member_p : sparseset_bit_p
* add_member : sparseset_set_bit
* remove_member : sparseset_clear_bit
* choose_one : sparseset_pop
Additionally, the sparse set representation supports enumeration of
the members in O(N) time, where n is the number of members in the set.
The members of the set are stored cache-friendly in the dense vector.
This makes it a competitive choice for iterating over relatively sparse
sets requiring operations:
* forall : EXECUTE_IF_SET_IN_SPARSESET
* set_copy : sparseset_copy
* set_intersection : sparseset_and
* set_union : sparseset_ior
* set_difference : sparseset_and_compl
* set_disjuction : (not implemented)
* set_compare : sparseset_equal_p
NB: It is OK to use remove_member during EXECUTE_IF_SET_IN_SPARSESET.
The iterator is updated for it.
Based on the efficiency of these operations, this representation of
sparse sets will often be superior to alternatives such as simple
bitmaps, linked-list bitmaps, array bitmaps, balanced binary trees,
hash tables, linked lists, etc., if the set is sufficiently sparse.
In the LOPLAS paper the cut-off point where sparse sets became faster
than simple bitmaps (see sbitmap.h) when N / U < 64 (where U is the
size of the universe of the set).
Because the set universe is fixed, the set cannot be resized. For
sparse sets with initially unknown size, linked-list bitmaps are a
better choice, see bitmap.h.
Sparse sets storage requirements are relatively large: O(U) with a
larger constant than sbitmaps (if the storage requirement for an
sbitmap with universe U is S, then the storage required for a sparse
set for the same universe are 2 * sizeof (SPARSESET_ELT_TYPE) * 8 * S).
Accessing the sparse vector is not very cache-friendly, but iterating
over the members in the set is cache-friendly because only the dense
vector is used. */
/* Data Structure used for the SparseSet representation. */
#define SPARSESET_ELT_TYPE unsigned int
typedef struct sparseset_def
{
SPARSESET_ELT_TYPE *dense; /* Dense array. */
SPARSESET_ELT_TYPE *sparse; /* Sparse array. */
SPARSESET_ELT_TYPE members; /* Number of elements. */
SPARSESET_ELT_TYPE size; /* Maximum number of elements. */
SPARSESET_ELT_TYPE iter; /* Iterator index. */
unsigned char iter_inc; /* Iteration increment amount. */
bool iterating;
SPARSESET_ELT_TYPE elms[2]; /* Combined dense and sparse arrays. */
} *sparseset;
#define sparseset_free(MAP) free(MAP)
extern sparseset sparseset_alloc (SPARSESET_ELT_TYPE n_elms);
extern void sparseset_clear_bit (sparseset, SPARSESET_ELT_TYPE);
extern void sparseset_copy (sparseset, sparseset);
extern void sparseset_and (sparseset, sparseset, sparseset);
extern void sparseset_and_compl (sparseset, sparseset, sparseset);
extern void sparseset_ior (sparseset, sparseset, sparseset);
extern bool sparseset_equal_p (sparseset, sparseset);
/* Operation: S = {}
Clear the set of all elements. */
inline void
sparseset_clear (sparseset s)
{
s->members = 0;
s->iterating = false;
}
/* Return the number of elements currently in the set. */
inline SPARSESET_ELT_TYPE
sparseset_cardinality (sparseset s)
{
return s->members;
}
/* Return the maximum number of elements this set can hold. */
inline SPARSESET_ELT_TYPE
sparseset_size (sparseset s)
{
return s->size;
}
/* Return true if e is a member of the set S, otherwise return false. */
inline bool
sparseset_bit_p (sparseset s, SPARSESET_ELT_TYPE e)
{
SPARSESET_ELT_TYPE idx;
gcc_checking_assert (e < s->size);
idx = s->sparse[e];
return idx < s->members && s->dense[idx] == e;
}
/* Low level insertion routine not meant for use outside of sparseset.[ch].
Assumes E is valid and not already a member of the set S. */
inline void
sparseset_insert_bit (sparseset s, SPARSESET_ELT_TYPE e, SPARSESET_ELT_TYPE idx)
{
s->sparse[e] = idx;
s->dense[idx] = e;
}
/* Operation: S = S + {e}
Insert E into the set S, if it isn't already a member. */
inline void
sparseset_set_bit (sparseset s, SPARSESET_ELT_TYPE e)
{
if (!sparseset_bit_p (s, e))
sparseset_insert_bit (s, e, s->members++);
}
/* Return and remove the last member added to the set S. */
inline SPARSESET_ELT_TYPE
sparseset_pop (sparseset s)
{
SPARSESET_ELT_TYPE mem = s->members;
gcc_checking_assert (mem != 0);
s->members = mem - 1;
return s->dense[s->members];
}
inline void
sparseset_iter_init (sparseset s)
{
s->iter = 0;
s->iter_inc = 1;
s->iterating = true;
}
inline bool
sparseset_iter_p (sparseset s)
{
if (s->iterating && s->iter < s->members)
return true;
else
return s->iterating = false;
}
inline SPARSESET_ELT_TYPE
sparseset_iter_elm (sparseset s)
{
return s->dense[s->iter];
}
inline void
sparseset_iter_next (sparseset s)
{
s->iter += s->iter_inc;
s->iter_inc = 1;
}
#define EXECUTE_IF_SET_IN_SPARSESET(SPARSESET, ITER) \
for (sparseset_iter_init (SPARSESET); \
sparseset_iter_p (SPARSESET) \
&& (((ITER) = sparseset_iter_elm (SPARSESET)) || 1); \
sparseset_iter_next (SPARSESET))
#endif /* GCC_SPARSESET_H */
|
