Mercurial > hg > truffle
annotate src/share/vm/utilities/hashtable.cpp @ 6162:e9140bf80b4a
7158800: Improve storage of symbol tables
Summary: Use an alternate version of hashing algorithm for symbol string tables and after a certain bucket size to improve performance
Reviewed-by: pbk, kamg, dlong, kvn, fparain
author | coleenp |
---|---|
date | Wed, 13 Jun 2012 19:52:59 -0400 |
parents | 436b4a3231bf |
children | 246d977b51f2 |
rev | line source |
---|---|
0 | 1 /* |
6162 | 2 * Copyright (c) 2003, 2012, Oracle and/or its affiliates. All rights reserved. |
0 | 3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. |
4 * | |
5 * This code is free software; you can redistribute it and/or modify it | |
6 * under the terms of the GNU General Public License version 2 only, as | |
7 * published by the Free Software Foundation. | |
8 * | |
9 * This code is distributed in the hope that it will be useful, but WITHOUT | |
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License | |
12 * version 2 for more details (a copy is included in the LICENSE file that | |
13 * accompanied this code). | |
14 * | |
15 * You should have received a copy of the GNU General Public License version | |
16 * 2 along with this work; if not, write to the Free Software Foundation, | |
17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. | |
18 * | |
1552
c18cbe5936b8
6941466: Oracle rebranding changes for Hotspot repositories
trims
parents:
470
diff
changeset
|
19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA |
c18cbe5936b8
6941466: Oracle rebranding changes for Hotspot repositories
trims
parents:
470
diff
changeset
|
20 * or visit www.oracle.com if you need additional information or have any |
c18cbe5936b8
6941466: Oracle rebranding changes for Hotspot repositories
trims
parents:
470
diff
changeset
|
21 * questions. |
0 | 22 * |
23 */ | |
24 | |
1972 | 25 #include "precompiled.hpp" |
26 #include "memory/allocation.inline.hpp" | |
27 #include "memory/resourceArea.hpp" | |
28 #include "oops/oop.inline.hpp" | |
29 #include "runtime/safepoint.hpp" | |
30 #include "utilities/dtrace.hpp" | |
31 #include "utilities/hashtable.hpp" | |
32 #include "utilities/hashtable.inline.hpp" | |
0 | 33 |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
34 |
4006 | 35 #ifndef USDT2 |
0 | 36 HS_DTRACE_PROBE_DECL4(hs_private, hashtable__new_entry, |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
37 void*, unsigned int, void*, void*); |
4006 | 38 #endif /* !USDT2 */ |
0 | 39 |
40 // This is a generic hashtable, designed to be used for the symbol | |
41 // and string tables. | |
42 // | |
43 // It is implemented as an open hash table with a fixed number of buckets. | |
44 // | |
45 // %note: | |
46 // - HashtableEntrys are allocated in blocks to reduce the space overhead. | |
47 | |
48 BasicHashtableEntry* BasicHashtable::new_entry(unsigned int hashValue) { | |
49 BasicHashtableEntry* entry; | |
50 | |
51 if (_free_list) { | |
52 entry = _free_list; | |
53 _free_list = _free_list->next(); | |
54 } else { | |
432 | 55 if (_first_free_entry + _entry_size >= _end_block) { |
56 int block_size = MIN2(512, MAX2((int)_table_size / 2, (int)_number_of_entries)); | |
0 | 57 int len = _entry_size * block_size; |
432 | 58 len = 1 << log2_intptr(len); // round down to power of 2 |
59 assert(len >= _entry_size, ""); | |
0 | 60 _first_free_entry = NEW_C_HEAP_ARRAY(char, len); |
61 _end_block = _first_free_entry + len; | |
62 } | |
63 entry = (BasicHashtableEntry*)_first_free_entry; | |
64 _first_free_entry += _entry_size; | |
65 } | |
66 | |
432 | 67 assert(_entry_size % HeapWordSize == 0, ""); |
0 | 68 entry->set_hash(hashValue); |
69 return entry; | |
70 } | |
71 | |
72 | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
73 template <class T> HashtableEntry<T>* Hashtable<T>::new_entry(unsigned int hashValue, T obj) { |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
74 HashtableEntry<T>* entry; |
0 | 75 |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
76 entry = (HashtableEntry<T>*)BasicHashtable::new_entry(hashValue); |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
77 entry->set_literal(obj); |
4006 | 78 #ifndef USDT2 |
0 | 79 HS_DTRACE_PROBE4(hs_private, hashtable__new_entry, |
80 this, hashValue, obj, entry); | |
4006 | 81 #else /* USDT2 */ |
82 HS_PRIVATE_HASHTABLE_NEW_ENTRY( | |
83 this, hashValue, (uintptr_t) obj, entry); | |
84 #endif /* USDT2 */ | |
0 | 85 return entry; |
86 } | |
87 | |
88 | |
6162 | 89 // Check to see if the hashtable is unbalanced. The caller set a flag to |
90 // rehash at the next safepoint. If this bucket is 60 times greater than the | |
91 // expected average bucket length, it's an unbalanced hashtable. | |
92 // This is somewhat an arbitrary heuristic but if one bucket gets to | |
93 // rehash_count which is currently 100, there's probably something wrong. | |
94 | |
95 bool BasicHashtable::check_rehash_table(int count) { | |
96 assert(table_size() != 0, "underflow"); | |
97 if (count > (((double)number_of_entries()/(double)table_size())*rehash_multiple)) { | |
98 // Set a flag for the next safepoint, which should be at some guaranteed | |
99 // safepoint interval. | |
100 return true; | |
101 } | |
102 return false; | |
103 } | |
104 | |
105 // Create a new table and using alternate hash code, populate the new table | |
106 // with the existing elements. This can be used to change the hash code | |
107 // and could in the future change the size of the table. | |
108 | |
109 template <class T> void Hashtable<T>::move_to(Hashtable<T>* new_table) { | |
110 int saved_entry_count = number_of_entries(); | |
111 | |
112 // Iterate through the table and create a new entry for the new table | |
113 for (int i = 0; i < new_table->table_size(); ++i) { | |
114 for (HashtableEntry<T>* p = bucket(i); p != NULL; ) { | |
115 HashtableEntry<T>* next = p->next(); | |
116 T string = p->literal(); | |
117 // Use alternate hashing algorithm on the symbol in the first table | |
118 unsigned int hashValue = new_hash(string); | |
119 // Get a new index relative to the new table (can also change size) | |
120 int index = new_table->hash_to_index(hashValue); | |
121 p->set_hash(hashValue); | |
122 unlink_entry(p); | |
123 new_table->add_entry(index, p); | |
124 p = next; | |
125 } | |
126 } | |
127 // give the new table the free list as well | |
128 new_table->copy_freelist(this); | |
129 assert(new_table->number_of_entries() == saved_entry_count, "lost entry on dictionary copy?"); | |
130 | |
131 // Destroy memory used by the buckets in the hashtable. The memory | |
132 // for the elements has been used in a new table and is not | |
133 // destroyed. The memory reuse will benefit resizing the SystemDictionary | |
134 // to avoid a memory allocation spike at safepoint. | |
135 free_buckets(); | |
136 } | |
137 | |
0 | 138 // Reverse the order of elements in the hash buckets. |
139 | |
140 void BasicHashtable::reverse() { | |
141 | |
142 for (int i = 0; i < _table_size; ++i) { | |
143 BasicHashtableEntry* new_list = NULL; | |
144 BasicHashtableEntry* p = bucket(i); | |
145 while (p != NULL) { | |
146 BasicHashtableEntry* next = p->next(); | |
147 p->set_next(new_list); | |
148 new_list = p; | |
149 p = next; | |
150 } | |
151 *bucket_addr(i) = new_list; | |
152 } | |
153 } | |
154 | |
155 | |
156 // Copy the table to the shared space. | |
157 | |
158 void BasicHashtable::copy_table(char** top, char* end) { | |
159 | |
160 // Dump the hash table entries. | |
161 | |
162 intptr_t *plen = (intptr_t*)(*top); | |
163 *top += sizeof(*plen); | |
164 | |
165 int i; | |
166 for (i = 0; i < _table_size; ++i) { | |
167 for (BasicHashtableEntry** p = _buckets[i].entry_addr(); | |
168 *p != NULL; | |
169 p = (*p)->next_addr()) { | |
170 if (*top + entry_size() > end) { | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
171 report_out_of_shared_space(SharedMiscData); |
0 | 172 } |
173 *p = (BasicHashtableEntry*)memcpy(*top, *p, entry_size()); | |
174 *top += entry_size(); | |
175 } | |
176 } | |
177 *plen = (char*)(*top) - (char*)plen - sizeof(*plen); | |
178 | |
179 // Set the shared bit. | |
180 | |
181 for (i = 0; i < _table_size; ++i) { | |
182 for (BasicHashtableEntry* p = bucket(i); p != NULL; p = p->next()) { | |
183 p->set_shared(); | |
184 } | |
185 } | |
186 } | |
187 | |
188 | |
189 | |
190 // Reverse the order of elements in the hash buckets. | |
191 | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
192 template <class T> void Hashtable<T>::reverse(void* boundary) { |
0 | 193 |
194 for (int i = 0; i < table_size(); ++i) { | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
195 HashtableEntry<T>* high_list = NULL; |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
196 HashtableEntry<T>* low_list = NULL; |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
197 HashtableEntry<T>* last_low_entry = NULL; |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
198 HashtableEntry<T>* p = bucket(i); |
0 | 199 while (p != NULL) { |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
200 HashtableEntry<T>* next = p->next(); |
0 | 201 if ((void*)p->literal() >= boundary) { |
202 p->set_next(high_list); | |
203 high_list = p; | |
204 } else { | |
205 p->set_next(low_list); | |
206 low_list = p; | |
207 if (last_low_entry == NULL) { | |
208 last_low_entry = p; | |
209 } | |
210 } | |
211 p = next; | |
212 } | |
213 if (low_list != NULL) { | |
214 *bucket_addr(i) = low_list; | |
215 last_low_entry->set_next(high_list); | |
216 } else { | |
217 *bucket_addr(i) = high_list; | |
218 } | |
219 } | |
220 } | |
221 | |
222 | |
223 // Dump the hash table buckets. | |
224 | |
225 void BasicHashtable::copy_buckets(char** top, char* end) { | |
226 intptr_t len = _table_size * sizeof(HashtableBucket); | |
227 *(intptr_t*)(*top) = len; | |
228 *top += sizeof(intptr_t); | |
229 | |
230 *(intptr_t*)(*top) = _number_of_entries; | |
231 *top += sizeof(intptr_t); | |
232 | |
233 if (*top + len > end) { | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
234 report_out_of_shared_space(SharedMiscData); |
0 | 235 } |
236 _buckets = (HashtableBucket*)memcpy(*top, _buckets, len); | |
237 *top += len; | |
238 } | |
239 | |
240 | |
241 #ifndef PRODUCT | |
242 | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
243 template <class T> void Hashtable<T>::print() { |
0 | 244 ResourceMark rm; |
245 | |
246 for (int i = 0; i < table_size(); i++) { | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
247 HashtableEntry<T>* entry = bucket(i); |
0 | 248 while(entry != NULL) { |
249 tty->print("%d : ", i); | |
250 entry->literal()->print(); | |
251 tty->cr(); | |
252 entry = entry->next(); | |
253 } | |
254 } | |
255 } | |
256 | |
257 | |
258 void BasicHashtable::verify() { | |
259 int count = 0; | |
260 for (int i = 0; i < table_size(); i++) { | |
261 for (BasicHashtableEntry* p = bucket(i); p != NULL; p = p->next()) { | |
262 ++count; | |
263 } | |
264 } | |
265 assert(count == number_of_entries(), "number of hashtable entries incorrect"); | |
266 } | |
267 | |
268 | |
269 #endif // PRODUCT | |
270 | |
271 | |
272 #ifdef ASSERT | |
273 | |
274 void BasicHashtable::verify_lookup_length(double load) { | |
275 if ((double)_lookup_length / (double)_lookup_count > load * 2.0) { | |
276 warning("Performance bug: SystemDictionary lookup_count=%d " | |
277 "lookup_length=%d average=%lf load=%f", | |
278 _lookup_count, _lookup_length, | |
279 (double) _lookup_length / _lookup_count, load); | |
280 } | |
281 } | |
282 | |
283 #endif | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
284 |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
285 // Explicitly instantiate these types |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
286 template class Hashtable<constantPoolOop>; |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
287 template class Hashtable<Symbol*>; |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
288 template class Hashtable<klassOop>; |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
289 template class Hashtable<oop>; |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
1972
diff
changeset
|
290 |