|
1 /* |
|
2 * Copyright (c) 2016, 2019, Oracle and/or its affiliates. All rights reserved. |
|
3 * ORACLE PROPRIETARY/CONFIDENTIAL. Use is subject to license terms. |
|
4 * |
|
5 * |
|
6 * |
|
7 * |
|
8 * |
|
9 * |
|
10 * |
|
11 * |
|
12 * |
|
13 * |
|
14 * |
|
15 * |
|
16 * |
|
17 * |
|
18 * |
|
19 * |
|
20 * |
|
21 * |
|
22 * |
|
23 * |
|
24 */ |
|
25 // package java.util; |
|
26 |
|
27 import java.math.BigInteger; |
|
28 import java.util.stream.Stream; |
|
29 |
|
30 /** |
|
31 * This interface is designed to provide a common protocol for objects |
|
32 * that generate pseudorandom sequences of numbers (or Boolean values) |
|
33 * and furthermore can easily <it>jump</it> forward (by a fixed amount) |
|
34 * to a distant point in the state cycle. |
|
35 * |
|
36 * <p>Ideally, all {@code JumpableRng} objects produced by iterative |
|
37 * jumping from a single original {@code JumpableRng} object are |
|
38 * statistically independent of one another and individually uniform. |
|
39 * In practice, one must settle for some approximation to independence |
|
40 * and uniformity. In particular, a specific implementation may |
|
41 * assume that each generator in a stream produced by the {@code jumps} |
|
42 * method is used to produce a number of values no larger than either |
|
43 * 2<sup>64</sup> or the square root of its period. Implementors are |
|
44 * advised to use algorithms whose period is at least 2<sup>127</sup>. |
|
45 * |
|
46 * <p>Methods are provided to perform a single jump operation and also |
|
47 * to produce a stream of generators produced from the original by |
|
48 * iterative copying and jumping of internal state. A typical |
|
49 * strategy for a multithreaded application is to create a single |
|
50 * {@code JumpableRng} object, calls its {@code jumps} method exactly |
|
51 * once, and then parcel out generators from the resulting stream, one |
|
52 * to each thread. It is generally not a good idea to call {@code jump} |
|
53 * on a generator that was itself produced by the {@code jumps} method, |
|
54 * because the result may be a generator identical to another |
|
55 * generator already produce by that call to the {@code jumps} method. |
|
56 * For this reason, the return type of the {@code jumps} method is |
|
57 * {@code Stream<Rng>} rather than {@code Stream<JumpableRng>}, even |
|
58 * though the actual generator objects in that stream likely do also |
|
59 * implement the {@code JumpableRng} interface. |
|
60 * |
|
61 * <p>An implementation of the {@code JumpableRng} interface must provide |
|
62 * concrete definitions for the methods {@code nextInt()}, {@code nextLong}, |
|
63 * {@code period()}, {@code copy()}, {@code jump()}, and {@code defaultJumpDistance()}. |
|
64 * Default implementations are provided for all other methods. |
|
65 * |
|
66 * <p>Objects that implement {@code java.util.JumpableRng} are |
|
67 * typically not cryptographically secure. Consider instead using |
|
68 * {@link java.security.SecureRandom} to get a cryptographically |
|
69 * secure pseudo-random number generator for use by |
|
70 * security-sensitive applications. |
|
71 * |
|
72 * @author Guy Steele |
|
73 * @since 1.9 |
|
74 */ |
|
75 interface JumpableRng extends StreamableRng { |
|
76 /** |
|
77 * Returns a new generator whose internal state is an exact copy |
|
78 * of this generator (therefore their future behavior should be |
|
79 * identical if subjected to the same series of operations). |
|
80 * |
|
81 * @return a new object that is a copy of this generator |
|
82 */ |
|
83 JumpableRng copy(); |
|
84 |
|
85 /** |
|
86 * Alter the state of this pseudorandom number generator so as to |
|
87 * jump forward a large, fixed distance (typically 2<sup>64</sup> |
|
88 * or more) within its state cycle. |
|
89 */ |
|
90 void jump(); |
|
91 |
|
92 /** |
|
93 * Returns the distance by which the {@code jump()} method will jump |
|
94 * forward within the state cycle of this generator object. |
|
95 * |
|
96 * @return the default jump distance (as a {@code double} value) |
|
97 */ |
|
98 double defaultJumpDistance(); |
|
99 |
|
100 /** |
|
101 * Returns an effectively unlimited stream of new pseudorandom |
|
102 * number generators, each of which implements the {@code Rng} |
|
103 * interface. |
|
104 * |
|
105 * @implNote It is permitted to implement this method in a manner |
|
106 * equivalent to {@code jumps(Long.MAX_VALUE)}. |
|
107 * |
|
108 * @implNote The default implementation produces a sequential stream |
|
109 * that repeatedly calls {@code copy()} and {@code jump()} on this generator, |
|
110 * and the copies become the generators produced by the stream. |
|
111 * |
|
112 * @return a stream of objects that implement the {@code Rng} interface |
|
113 */ |
|
114 default Stream<Rng> jumps() { |
|
115 return Stream.generate(this::copyAndJump).sequential(); |
|
116 } |
|
117 |
|
118 /** |
|
119 * Returns a stream producing the given {@code streamSize} number of |
|
120 * new pseudorandom number generators, each of which implements the |
|
121 * {@code Rng} interface. |
|
122 * |
|
123 * @implNote The default implementation produces a sequential stream |
|
124 * that repeatedly calls {@code copy()} and {@code jump()} on this generator, |
|
125 * and the copies become the generators produced by the stream. |
|
126 * |
|
127 * @param streamSize the number of generators to generate |
|
128 * @return a stream of objects that implement the {@code Rng} interface |
|
129 * @throws IllegalArgumentException if {@code streamSize} is |
|
130 * less than zero |
|
131 */ |
|
132 default Stream<Rng> jumps(long streamSize) { |
|
133 return jumps().limit(streamSize); |
|
134 } |
|
135 |
|
136 /** |
|
137 * Returns an effectively unlimited stream of new pseudorandom |
|
138 * number generators, each of which implements the {@code Rng} |
|
139 * interface. Ideally the generators in the stream will appear |
|
140 * to be statistically independent. |
|
141 * |
|
142 * @implNote The default implementation calls {@code jumps()}. |
|
143 * |
|
144 * @return a stream of objects that implement the {@code Rng} interface |
|
145 */ |
|
146 default Stream<Rng> rngs() { |
|
147 return this.jumps(); |
|
148 } |
|
149 |
|
150 /** |
|
151 * Returns a stream producing the given {@code streamSize} number of |
|
152 * new pseudorandom number generators, each of which implements the |
|
153 * {@code Rng} interface. Ideally the generators in the stream will |
|
154 * appear to be statistically independent. |
|
155 * |
|
156 * @implNote The default implementation calls {@code jumps(streamSize)}. |
|
157 * |
|
158 * @param streamSize the number of generators to generate |
|
159 * @return a stream of objects that implement the {@code Rng} interface |
|
160 * @throws IllegalArgumentException if {@code streamSize} is |
|
161 * less than zero |
|
162 */ |
|
163 default Stream<Rng> rngs(long streamSize) { |
|
164 return this.jumps(streamSize); |
|
165 } |
|
166 |
|
167 /** |
|
168 * Copy this generator, jump this generator forward, then return the copy. |
|
169 */ |
|
170 default Rng copyAndJump() { |
|
171 Rng result = copy(); |
|
172 jump(); |
|
173 return result; |
|
174 } |
|
175 |
|
176 } |