1DRAND48(P) POSIX Programmer's Manual DRAND48(P)
2
3
4
6 drand48, erand48, jrand48, lcong48, lrand48, mrand48, nrand48, seed48,
7 srand48 - generate uniformly distributed pseudo-random numbers
8
10 #include <stdlib.h>
11
12 double drand48(void);
13 double erand48(unsigned short xsubi[3]);
14 long jrand48(unsigned short xsubi[3]);
15 void lcong48(unsigned short param[7]);
16 long lrand48(void);
17 long mrand48(void);
18 long nrand48(unsigned short xsubi[3]);
19 unsigned short *seed48(unsigned short seed16v[3]);
20 void srand48(long seedval);
21
22
24 This family of functions shall generate pseudo-random numbers using a
25 linear congruential algorithm and 48-bit integer arithmetic.
26
27 The drand48() and erand48() functions shall return non-negative, dou‐
28 ble-precision, floating-point values, uniformly distributed over the
29 interval [0.0,1.0).
30
31 The lrand48() and nrand48() functions shall return non-negative, long
32 integers, uniformly distributed over the interval [0,2**31).
33
34 The mrand48() and jrand48() functions shall return signed long integers
35 uniformly distributed over the interval [-2**31,2**31).
36
37 The srand48(), seed48(), and lcong48() functions are initialization
38 entry points, one of which should be invoked before either drand48(),
39 lrand48(), or mrand48() is called. (Although it is not recommended
40 practice, constant default initializer values shall be supplied auto‐
41 matically if drand48(), lrand48(), or mrand48() is called without a
42 prior call to an initialization entry point.) The erand48(), nrand48(),
43 and jrand48() functions do not require an initialization entry point to
44 be called first.
45
46 All the routines work by generating a sequence of 48-bit integer val‐
47 ues, X_i , according to the linear congruential formula: X_n+1 = (aX_n
48 + c)_mod m n>= 0
49
50 The parameter m = 2**48; hence 48-bit integer arithmetic is performed.
51 Unless lcong48() is invoked, the multiplier value a and the addend
52 value c are given by: a = 5DEECE66D_16 = 273673163155_8
53
54 c = B_16 = 13_8
55
56 The value returned by any of the drand48(), erand48(), jrand48(),
57 lrand48(), mrand48(), or nrand48() functions is computed by first gen‐
58 erating the next 48-bit X_i in the sequence. Then the appropriate num‐
59 ber of bits, according to the type of data item to be returned, are
60 copied from the high-order (leftmost) bits of X_i and transformed into
61 the returned value.
62
63 The drand48(), lrand48(), and mrand48() functions store the last 48-bit
64 X_i generated in an internal buffer; that is why the application shall
65 ensure that these are initialized prior to being invoked. The
66 erand48(), nrand48(), and jrand48() functions require the calling pro‐
67 gram to provide storage for the successive X_i values in the array
68 specified as an argument when the functions are invoked. That is why
69 these routines do not have to be initialized; the calling program
70 merely has to place the desired initial value of X_i into the array and
71 pass it as an argument. By using different arguments, erand48(),
72 nrand48(), and jrand48() allow separate modules of a large program to
73 generate several independent streams of pseudo-random numbers; that is,
74 the sequence of numbers in each stream shall not depend upon how many
75 times the routines are called to generate numbers for the other
76 streams.
77
78 The initializer function srand48() sets the high-order 32 bits of X_i
79 to the low-order 32 bits contained in its argument. The low-order 16
80 bits of X_i are set to the arbitrary value 330E_16.
81
82 The initializer function seed48() sets the value of X_i to the 48-bit
83 value specified in the argument array. The low-order 16 bits of X_i are
84 set to the low-order 16 bits of seed16v[0]. The mid-order 16 bits of
85 X_i are set to the low-order 16 bits of seed16v[1]. The high-order 16
86 bits of X_i are set to the low-order 16 bits of seed16v[2]. In addi‐
87 tion, the previous value of X_i is copied into a 48-bit internal buf‐
88 fer, used only by seed48(), and a pointer to this buffer is the value
89 returned by seed48(). This returned pointer, which can just be ignored
90 if not needed, is useful if a program is to be restarted from a given
91 point at some future time-use the pointer to get at and store the last
92 X_i value, and then use this value to reinitialize via seed48() when
93 the program is restarted.
94
95 The initializer function lcong48() allows the user to specify the ini‐
96 tial X_i , the multiplier value a, and the addend value c. Argument
97 array elements param[0-2] specify X_i , param[3-5] specify the multi‐
98 plier a, and param[6] specifies the 16-bit addend c. After lcong48() is
99 called, a subsequent call to either srand48() or seed48() shall restore
100 the standard multiplier and addend values, a and c, specified above.
101
102 The drand48(), lrand48(), and mrand48() functions need not be reen‐
103 trant. A function that is not required to be reentrant is not required
104 to be thread-safe.
105
107 As described in the DESCRIPTION above.
108
110 No errors are defined.
111
112 The following sections are informative.
113
115 None.
116
118 None.
119
121 None.
122
124 None.
125
127 rand() , the Base Definitions volume of IEEE Std 1003.1-2001,
128 <stdlib.h>
129
131 Portions of this text are reprinted and reproduced in electronic form
132 from IEEE Std 1003.1, 2003 Edition, Standard for Information Technology
133 -- Portable Operating System Interface (POSIX), The Open Group Base
134 Specifications Issue 6, Copyright (C) 2001-2003 by the Institute of
135 Electrical and Electronics Engineers, Inc and The Open Group. In the
136 event of any discrepancy between this version and the original IEEE and
137 The Open Group Standard, the original IEEE and The Open Group Standard
138 is the referee document. The original Standard can be obtained online
139 at http://www.opengroup.org/unix/online.html .
140
141
142
143IEEE/The Open Group 2003 DRAND48(P)