5 |
|
*/ |
6 |
|
|
7 |
|
package jsr166e; |
8 |
– |
import java.util.Arrays; |
8 |
|
import java.util.Random; |
9 |
|
import java.util.concurrent.atomic.AtomicInteger; |
10 |
|
import java.util.concurrent.atomic.AtomicLong; |
45 |
|
* A StripedAdder maintains a table of Atomic long variables. The |
46 |
|
* table is indexed by per-thread hash codes. |
47 |
|
* |
48 |
< |
* By default, the table is lazily initialized, to minimize |
49 |
< |
* footprint until adders are used. On first use, the table is set |
50 |
< |
* to size DEFAULT_INITIAL_SIZE (currently 8). Table size is |
51 |
< |
* bounded by the number of CPUS (if larger than the default |
52 |
< |
* size). |
48 |
> |
* Table entries are of class Adder; a variant of AtomicLong |
49 |
> |
* padded to reduce cache contention on most processors. Padding |
50 |
> |
* is overkill for most Atomics because they are usually |
51 |
> |
* irregularly scattered in memory and thus don't interfere much |
52 |
> |
* with each other. But Atomic objects residing in arrays will |
53 |
> |
* tend to be placed adjacent to each other, and so will most |
54 |
> |
* often share cache lines (with a huge negative performance |
55 |
> |
* impact) without this precaution. |
56 |
> |
* |
57 |
> |
* Because Adders are relatively large, we avoid creating them |
58 |
> |
* until they are needed. On the other hand, we try to create them |
59 |
> |
* on any sign of contention. |
60 |
|
* |
61 |
|
* Per-thread hash codes are initialized to random values. |
62 |
|
* Collisions are indicated by failed CASes when performing an add |
68 |
|
* hashing", using a secondary hash (Marsaglia XorShift) to try to |
69 |
|
* find a free slot. |
70 |
|
* |
71 |
< |
* The table size is capped because, when there are more threads |
72 |
< |
* than CPUs, supposing that each thread were bound to a CPU, |
73 |
< |
* there would exist a perfect hash function mapping threads to |
74 |
< |
* slots that eliminates collisions. When we reach capacity, we |
75 |
< |
* search for this mapping by randomly varying the hash codes of |
71 |
> |
* By default, the table is lazily initialized. Upon first use, |
72 |
> |
* the table is set to size 2 (the minimum non-empty size), but |
73 |
> |
* containing only a single Adder. The maximum table size is |
74 |
> |
* bounded by nearest power of two >= the number of CPUS. The |
75 |
> |
* table size is capped because, when there are more threads than |
76 |
> |
* CPUs, supposing that each thread were bound to a CPU, there |
77 |
> |
* would exist a perfect hash function mapping threads to slots |
78 |
> |
* that eliminates collisions. When we reach capacity, we search |
79 |
> |
* for this mapping by randomly varying the hash codes of |
80 |
|
* colliding threads. Because search is random, and failures only |
81 |
|
* become known via CAS failures, convergence will be slow, and |
82 |
|
* because threads are typically not bound to CPUS forever, may |
83 |
|
* not occur at all. However, despite these limitations, observed |
84 |
|
* contention is typically low in these cases. |
85 |
|
* |
76 |
– |
* Table entries are of class Adder; a form of AtomicLong padded |
77 |
– |
* to reduce cache contention on most processors. Padding is |
78 |
– |
* overkill for most Atomics because they are usually irregularly |
79 |
– |
* scattered in memory and thus don't interfere much with each |
80 |
– |
* other. But Atomic objects residing in arrays will tend to be |
81 |
– |
* placed adjacent to each other, and so will most often share |
82 |
– |
* cache lines without this precaution. Adders are by default |
83 |
– |
* constructed upon first use, which further improves per-thread |
84 |
– |
* locality and helps reduce footprint. |
85 |
– |
* |
86 |
|
* A single spinlock is used for resizing the table as well as |
87 |
< |
* populating slots with new Adders. Upon lock contention, threads |
87 |
> |
* populating slots with new Adders. After initialization, there |
88 |
> |
* is no need for a blocking lock: Upon lock contention, threads |
89 |
|
* try other slots rather than blocking. After initialization, at |
90 |
|
* least one slot exists, so retries will eventually find a |
91 |
|
* candidate Adder. During these retries, there is increased |
93 |
|
* alternatives. |
94 |
|
*/ |
95 |
|
|
95 |
– |
/** |
96 |
– |
* Padded version of AtomicLong |
97 |
– |
*/ |
98 |
– |
static final class Adder extends AtomicLong { |
99 |
– |
long p0, p1, p2, p3, p4, p5, p6, p7, p8, p9, pa, pb, pc, pd, pe; |
100 |
– |
Adder(long x) { super(x); } |
101 |
– |
} |
102 |
– |
|
96 |
|
private static final int NCPU = Runtime.getRuntime().availableProcessors(); |
97 |
|
|
98 |
|
/** |
99 |
< |
* Table bounds. DEFAULT_INITIAL_SIZE is the table size set upon |
100 |
< |
* first use under default constructor, and must be a power of |
101 |
< |
* two. There is not much point in making size a lot smaller than |
102 |
< |
* that of Adders though. CAP is the maximum allowed table size. |
103 |
< |
*/ |
104 |
< |
private static final int DEFAULT_INITIAL_SIZE = 8; |
105 |
< |
private static final int CAP = Math.max(NCPU, DEFAULT_INITIAL_SIZE); |
99 |
> |
* Padded variant of AtomicLong. The value field is placed |
100 |
> |
* between pads, hoping that the JVM doesn't reorder them. |
101 |
> |
* Updates are via inlined CAS in methods add and retryAdd. |
102 |
> |
*/ |
103 |
> |
static final class Adder { |
104 |
> |
volatile long p0, p1, p2, p3, p4, p5, p6; |
105 |
> |
volatile long value; |
106 |
> |
volatile long q0, q1, q2, q3, q4, q5, q6; |
107 |
> |
Adder(long x) { value = x; } |
108 |
> |
} |
109 |
|
|
110 |
|
/** |
111 |
|
* Holder for the thread-local hash code. The code is initially |
115 |
|
static final Random rng = new Random(); |
116 |
|
int code; |
117 |
|
HashCode() { |
118 |
< |
int h = rng.nextInt(); |
119 |
< |
code = (h == 0) ? 1 : h; // ensure nonzero |
118 |
> |
int h = rng.nextInt(); // Avoid zero, because of xorShift rehash |
119 |
> |
code = (h == 0) ? 1 : h; |
120 |
|
} |
121 |
|
} |
122 |
|
|
136 |
|
static final ThreadHashCode threadHashCode = new ThreadHashCode(); |
137 |
|
|
138 |
|
/** |
139 |
< |
* Table of adders. Size is power of two, grows to be at most CAP. |
139 |
> |
* Table of adders. When non-null, size is a power of 2, at least 2. |
140 |
|
*/ |
141 |
|
private transient volatile Adder[] adders; |
142 |
|
|
143 |
|
/** |
144 |
< |
* Serves as a lock when resizing and/or creating Adders. There |
149 |
< |
* is no need for a blocking lock: Except during initialization |
150 |
< |
* races, when busy, other threads try other slots. However, |
151 |
< |
* during (double-checked) initializations, we use the |
152 |
< |
* "synchronized" lock on this object. |
144 |
> |
* Spinlock (locked via CAS) used when resizing and/or creating Adders. |
145 |
|
*/ |
146 |
< |
private final AtomicInteger mutex; |
146 |
> |
private volatile int busy; |
147 |
|
|
148 |
|
/** |
149 |
|
* Creates a new adder with zero sum. |
150 |
|
*/ |
151 |
|
public StripedAdder() { |
160 |
– |
this.mutex = new AtomicInteger(); |
161 |
– |
// remaining initialization on first call to add. |
152 |
|
} |
153 |
|
|
154 |
|
/** |
159 |
|
* will concurrently update the sum. |
160 |
|
*/ |
161 |
|
public StripedAdder(int expectedContention) { |
162 |
< |
int cap = (expectedContention < CAP) ? expectedContention : CAP; |
163 |
< |
int size = 1; |
162 |
> |
int cap = (expectedContention < NCPU) ? expectedContention : NCPU; |
163 |
> |
int size = 2; |
164 |
|
while (size < cap) |
165 |
|
size <<= 1; |
166 |
|
Adder[] as = new Adder[size]; |
167 |
|
for (int i = 0; i < size; ++i) |
168 |
|
as[i] = new Adder(0); |
169 |
|
this.adders = as; |
180 |
– |
this.mutex = new AtomicInteger(); |
170 |
|
} |
171 |
|
|
172 |
|
/** |
175 |
|
* @param x the value to add |
176 |
|
*/ |
177 |
|
public void add(long x) { |
178 |
< |
Adder[] as; Adder a; int n; long v; // locals to hold volatile reads |
178 |
> |
Adder[] as; Adder a; int n; // locals to hold volatile reads |
179 |
|
HashCode hc = threadHashCode.get(); |
180 |
|
int h = hc.code; |
181 |
< |
if ((as = adders) == null || (n = as.length) < 1 || |
182 |
< |
(a = as[(n - 1) & h]) == null || |
183 |
< |
!a.compareAndSet(v = a.get(), v + x)) |
184 |
< |
retryAdd(x, hc); |
181 |
> |
boolean collide; |
182 |
> |
if ((as = adders) != null && (n = as.length) > 0 && |
183 |
> |
(a = as[(n - 1) & h]) != null) { |
184 |
> |
long v = a.value; |
185 |
> |
if (UNSAFE.compareAndSwapLong(a, valueOffset, v, v + x)) |
186 |
> |
return; |
187 |
> |
collide = true; |
188 |
> |
} |
189 |
> |
else |
190 |
> |
collide = false; |
191 |
> |
retryAdd(x, hc, collide); |
192 |
|
} |
193 |
|
|
194 |
|
/** |
195 |
|
* Handle cases of add involving initialization, resizing, |
196 |
|
* creating new Adders, and/or contention. See above for |
197 |
< |
* explanation. |
197 |
> |
* explanation. This method suffers the usual non-modularity |
198 |
> |
* problems of optimistic retry code, relying on rechecked sets of |
199 |
> |
* reads. |
200 |
|
*/ |
201 |
< |
private void retryAdd(long x, HashCode hc) { |
201 |
> |
private void retryAdd(long x, HashCode hc, boolean collide) { |
202 |
|
int h = hc.code; |
205 |
– |
final AtomicInteger mutex = this.mutex; |
206 |
– |
int collisions = 1 - mutex.get(); // first guess: collides if not locked |
203 |
|
for (;;) { |
204 |
< |
Adder[] as; Adder a; long v; int k, n; |
205 |
< |
while ((as = adders) == null || (n = as.length) < 1) { |
206 |
< |
synchronized(mutex) { // Try to initialize |
207 |
< |
if (adders == null) { |
208 |
< |
Adder[] rs = new Adder[DEFAULT_INITIAL_SIZE]; |
209 |
< |
rs[h & (DEFAULT_INITIAL_SIZE - 1)] = new Adder(0); |
210 |
< |
adders = rs; |
204 |
> |
Adder[] as; Adder a; int n; |
205 |
> |
if ((as = adders) != null && (n = as.length) > 0) { |
206 |
> |
if ((a = as[(n - 1) & h]) != null) { |
207 |
> |
boolean shared = true; // Slot exists |
208 |
> |
if (collide && n < NCPU && busy == 0 && |
209 |
> |
UNSAFE.compareAndSwapInt(this, busyOffset, 0, 1)) { |
210 |
> |
try { |
211 |
> |
if (adders == as) { // Expand table |
212 |
> |
Adder[] rs = new Adder[n << 1]; |
213 |
> |
for (int i = 0; i < n; ++i) |
214 |
> |
rs[i] = as[i]; |
215 |
> |
adders = rs; |
216 |
> |
shared = false; |
217 |
> |
} |
218 |
> |
} finally { |
219 |
> |
busy = 0; |
220 |
> |
} |
221 |
> |
if (shared || (h & n) != 0) { |
222 |
> |
collide = false; |
223 |
> |
continue; // Array or index changed |
224 |
> |
} |
225 |
> |
} |
226 |
> |
long v = a.value; |
227 |
> |
if (UNSAFE.compareAndSwapLong(a, valueOffset, v, v + x)) |
228 |
> |
break; |
229 |
> |
collide = shared; |
230 |
> |
} |
231 |
> |
else { // Try to attach new Adder |
232 |
> |
if (busy == 0 && |
233 |
> |
UNSAFE.compareAndSwapInt(this, busyOffset, 0, 1)) { |
234 |
> |
boolean created = false; |
235 |
> |
try { // Recheck under lock |
236 |
> |
Adder[] rs; int m, j; |
237 |
> |
if ((rs = adders) != null && (m = rs.length) > 0 && |
238 |
> |
rs[j = (m - 1) & h] == null) { |
239 |
> |
rs[j] = new Adder(x); |
240 |
> |
created = true; |
241 |
> |
} |
242 |
> |
} finally { |
243 |
> |
busy = 0; |
244 |
> |
} |
245 |
> |
if (created) |
246 |
> |
break; |
247 |
> |
continue; // Slot is now non-empty |
248 |
|
} |
249 |
+ |
collide = false; |
250 |
|
} |
251 |
< |
collisions = 0; |
251 |
> |
h ^= h << 13; // Rehash |
252 |
> |
h ^= h >>> 17; |
253 |
> |
h ^= h << 5; |
254 |
|
} |
255 |
< |
|
256 |
< |
if ((a = as[k = (n - 1) & h]) == null) { // Try to add slot |
257 |
< |
if (mutex.get() == 0 && mutex.compareAndSet(0, 1)) { |
255 |
> |
else if (busy == 0) { // Default-initialize |
256 |
> |
Adder r = new Adder(x); |
257 |
> |
Adder[] rs = new Adder[2]; |
258 |
> |
rs[h & 1] = r; |
259 |
> |
if (adders == as && busy == 0 && |
260 |
> |
UNSAFE.compareAndSwapInt(this, busyOffset, 0, 1)) { |
261 |
> |
boolean init = false; |
262 |
|
try { |
263 |
< |
if (adders == as && as[k] == null) |
264 |
< |
a = as[k] = new Adder(x); |
263 |
> |
if (adders == as) { |
264 |
> |
adders = rs; |
265 |
> |
init = true; |
266 |
> |
} |
267 |
|
} finally { |
268 |
< |
mutex.set(0); |
268 |
> |
busy = 0; |
269 |
|
} |
270 |
< |
if (a != null) |
270 |
> |
if (init) |
271 |
|
break; |
272 |
|
} |
231 |
– |
collisions = 0; |
273 |
|
} |
274 |
< |
else if (collisions != 0 && n < CAP && // Try to expand table |
275 |
< |
mutex.get() == 0 && mutex.compareAndSet(0, 1)) { |
235 |
< |
try { |
236 |
< |
if (adders == as) { |
237 |
< |
Adder[] rs = new Adder[n << 1]; |
238 |
< |
for (int i = 0; i < n; ++i) |
239 |
< |
rs[i] = as[i]; |
240 |
< |
adders = rs; |
241 |
< |
} |
242 |
< |
} finally { |
243 |
< |
mutex.set(0); |
244 |
< |
} |
245 |
< |
collisions = 0; |
246 |
< |
} |
247 |
< |
else if (a.compareAndSet(v = a.get(), v + x)) |
248 |
< |
break; |
249 |
< |
else |
250 |
< |
collisions = 1; |
251 |
< |
h ^= h << 13; // Rehash |
252 |
< |
h ^= h >>> 17; |
253 |
< |
h ^= h << 5; |
274 |
> |
else if (adders == as) // Lost initialization race |
275 |
> |
Thread.yield(); |
276 |
|
} |
277 |
< |
hc.code = h; |
277 |
> |
hc.code = h; // Record index for next time |
278 |
|
} |
279 |
|
|
280 |
|
/** |
292 |
|
for (int i = 0; i < n; ++i) { |
293 |
|
Adder a = as[i]; |
294 |
|
if (a != null) |
295 |
< |
sum += a.get(); |
295 |
> |
sum += a.value; |
296 |
|
} |
297 |
|
} |
298 |
|
return sum; |
310 |
|
for (int i = 0; i < n; ++i) { |
311 |
|
Adder a = as[i]; |
312 |
|
if (a != null) |
313 |
< |
a.set(0L); |
313 |
> |
a.value = 0L; |
314 |
|
} |
315 |
|
} |
316 |
|
} |
342 |
|
for (int i = 0; i < n; ++i) { |
343 |
|
Adder a = as[i]; |
344 |
|
if (a != null) { |
345 |
< |
sum += a.get(); |
346 |
< |
a.set(0L); |
345 |
> |
sum += a.value; |
346 |
> |
a.value = 0L; |
347 |
|
} |
348 |
|
} |
349 |
|
} |
359 |
|
private void readObject(ObjectInputStream s) |
360 |
|
throws IOException, ClassNotFoundException { |
361 |
|
s.defaultReadObject(); |
362 |
< |
mutex.set(0); |
362 |
> |
busy = 0; |
363 |
|
add(s.readLong()); |
364 |
|
} |
365 |
|
|
366 |
+ |
// Unsafe mechanics |
367 |
+ |
private static final sun.misc.Unsafe UNSAFE; |
368 |
+ |
private static final long busyOffset; |
369 |
+ |
private static final long valueOffset; |
370 |
+ |
static { |
371 |
+ |
try { |
372 |
+ |
UNSAFE = getUnsafe(); |
373 |
+ |
Class<?> sk = StripedAdder.class; |
374 |
+ |
busyOffset = UNSAFE.objectFieldOffset |
375 |
+ |
(sk.getDeclaredField("busy")); |
376 |
+ |
Class<?> ak = Adder.class; |
377 |
+ |
valueOffset = UNSAFE.objectFieldOffset |
378 |
+ |
(ak.getDeclaredField("value")); |
379 |
+ |
} catch (Exception e) { |
380 |
+ |
throw new Error(e); |
381 |
+ |
} |
382 |
+ |
} |
383 |
+ |
|
384 |
+ |
/** |
385 |
+ |
* Returns a sun.misc.Unsafe. Suitable for use in a 3rd party package. |
386 |
+ |
* Replace with a simple call to Unsafe.getUnsafe when integrating |
387 |
+ |
* into a jdk. |
388 |
+ |
* |
389 |
+ |
* @return a sun.misc.Unsafe |
390 |
+ |
*/ |
391 |
+ |
private static sun.misc.Unsafe getUnsafe() { |
392 |
+ |
try { |
393 |
+ |
return sun.misc.Unsafe.getUnsafe(); |
394 |
+ |
} catch (SecurityException se) { |
395 |
+ |
try { |
396 |
+ |
return java.security.AccessController.doPrivileged |
397 |
+ |
(new java.security |
398 |
+ |
.PrivilegedExceptionAction<sun.misc.Unsafe>() { |
399 |
+ |
public sun.misc.Unsafe run() throws Exception { |
400 |
+ |
java.lang.reflect.Field f = sun.misc |
401 |
+ |
.Unsafe.class.getDeclaredField("theUnsafe"); |
402 |
+ |
f.setAccessible(true); |
403 |
+ |
return (sun.misc.Unsafe) f.get(null); |
404 |
+ |
}}); |
405 |
+ |
} catch (java.security.PrivilegedActionException e) { |
406 |
+ |
throw new RuntimeException("Could not initialize intrinsics", |
407 |
+ |
e.getCause()); |
408 |
+ |
} |
409 |
+ |
} |
410 |
+ |
} |
411 |
+ |
|
412 |
|
} |