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
|
// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#include "bitvectorcache.h"
#include <vespa/vespalib/stllike/hash_map.hpp>
#include <algorithm>
#include <cassert>
#include <cinttypes>
#include <mutex>
#include <vespa/log/log.h>
LOG_SETUP(".searchlib.common.bitvectorcache");
namespace search {
BitVectorCache::BitVectorCache(GenerationHolder &genHolder)
: _lookupCount(0),
_needPopulation(false),
_mutex(),
_keys(),
_chunks(),
_genHolder(genHolder)
{
}
BitVectorCache::~BitVectorCache() = default;
void
BitVectorCache::computeCountVector(KeySet & keys, CountVector & v) const
{
std::vector<Key> notFound;
std::vector<CondensedBitVector::KeySet> keySets;
ChunkV chunks;
{
std::shared_lock guard(_mutex);
keySets.resize(_chunks.size());
auto end = _keys.end();
for (Key k : keys) {
auto found = _keys.find(k);
if (found != end) {
const KeyMeta & m = found->second;
keySets[m.chunkId()].insert(m.chunkIndex());
}
}
chunks = _chunks;
}
for (Key k : notFound) {
keys.erase(k);
}
size_t index(0);
if (chunks.empty()) {
memset(&v[0], 0, v.size());
}
for (const auto & chunk : chunks) {
if (index == 0) {
chunk->initializeCountVector(keySets[index++], v);
} else {
chunk->addCountVector(keySets[index++], v);
}
}
}
BitVectorCache::KeySet
BitVectorCache::lookupCachedSet(const KeyAndCountSet & keys)
{
KeySet cached(keys.size()*3);
std::shared_lock shared_guard(_mutex);
uint64_t lookupCount = _lookupCount++;
if (lookupCount == 2000) {
requirePopulation();
} else if ((lookupCount & 0x1fffff) == 0x100000) {
if (hasCostChanged(shared_guard)) {
requirePopulation();
}
}
for (const auto & e : keys) {
auto found = _keys.find(e.first);
if (found != _keys.end()) {
KeyMeta & m = found->second;
m.lookup();
if (m.isCached()) {
cached.insert(e.first);
}
} else {
shared_guard.unlock();
{
std::unique_lock unique_guard(_mutex);
_keys[e.first] = KeyMeta().lookup().bitCount(e.second);
}
shared_guard.lock();
}
}
return cached;
}
BitVectorCache::SortedKeyMeta
BitVectorCache::getSorted(Key2Index & keys)
{
SortedKeyMeta sorted;
sorted.reserve(keys.size());
for (auto & e : keys) {
sorted.emplace_back(e.first, &e.second);
}
std::sort(sorted.begin(), sorted.end(),
[&] (const auto & a, const auto & b) {
return a.second->cost() > b.second->cost();
});
return sorted;
}
bool
BitVectorCache::hasCostChanged(const std::shared_lock<std::shared_mutex> & guard)
{
(void) guard;
if ( ! _chunks.empty()) {
SortedKeyMeta sorted(getSorted(_keys));
double oldCached(0);
for (auto & e : sorted) {
const KeyMeta & m = *e.second;
if ( m.isCached() ) {
oldCached += m.cost();
}
}
double newCached(0);
for (size_t i(0); i < sorted.size() && i < _chunks[0]->getKeyCapacity(); i++) {
const KeyMeta & m = *sorted[i].second;
newCached += m.cost();
}
if (newCached > oldCached * 1.01) { // 1% change needed.
return true;
}
}
return false;
}
void
BitVectorCache::populate(Key2Index & newKeys, CondensedBitVector & chunk, const PopulateInterface & lookup)
{
SortedKeyMeta sorted(getSorted(newKeys));
double sum(0);
for (auto & e : sorted) {
e.second->unCache();
sum += e.second->cost();
}
double accum(0.0);
uint32_t index(0);
for (const auto & e : sorted) {
KeyMeta & m = *e.second;
if (index >= chunk.getKeyCapacity()) {
assert( ! m.isCached());
} else {
double percentage(m.cost()*100.0/sum);
accum += percentage;
m.chunkId(0);
m.chunkIndex(index);
LOG(debug, "Populating bitvector %2d with feature %" PRIu64 " and %ld bits set. Cost is %8f = %2.2f%%, accumulated cost is %2.2f%%",
index, e.first, m.bitCount(), m.cost(), percentage, accum);
assert(m.isCached());
assert(newKeys[e.first].isCached());
assert(&m == &newKeys[e.first]);
PopulateInterface::Iterator::UP iterator = lookup.lookup(e.first);
if (iterator) {
for (int32_t docId(iterator->getNext()); docId >= 0; docId = iterator->getNext()) {
chunk.set(m.chunkIndex(), docId, true);
}
} else {
LOG(error, "Unable to to find a valid iterator for feature %" PRIu64 " and %ld bits set at while populating bitvector %2d. This should in theory be impossible.",
e.first, m.bitCount(), index);
}
index++;
}
}
}
void
BitVectorCache::populate(uint32_t sz, const PopulateInterface & lookup)
{
if (!needPopulation()) return;
std::unique_lock guard(_mutex);
Key2Index newKeys(_keys);
guard.unlock();
CondensedBitVector::UP chunk(CondensedBitVector::create(sz, _genHolder));
populate(newKeys, *chunk, lookup);
guard.lock();
_chunks.push_back(std::move(chunk));
_keys.swap(newKeys);
_needPopulation = false;
}
void
BitVectorCache::set(Key key, uint32_t index, bool v)
{
std::shared_lock guard(_mutex);
auto found = _keys.find(key);
if (found != _keys.end()) {
const KeyMeta & m(found->second);
if (m.isCached()) {
_chunks[m.chunkId()]->set(m.chunkIndex(), index, v);
}
}
}
bool
BitVectorCache::get(Key key, uint32_t index) const
{
(void) key; (void) index;
return false;
}
void
BitVectorCache::removeIndex(uint32_t index)
{
std::unique_lock guard(_mutex);
for (auto & chunk : _chunks) {
chunk->clearIndex(index);
}
}
void
BitVectorCache::adjustDocIdLimit(uint32_t docId)
{
for (auto &chunk : _chunks) {
chunk->adjustDocIdLimit(docId);
}
}
}
|
