| 1 | /* RAND functions using 64b INTEGERs | 
|---|
| 2 |  | 
|---|
| 3 |   F. CANTONNET - HPCL - GWU */ | 
|---|
| 4 |  | 
|---|
| 5 | double randlc (double *x, double a) | 
|---|
| 6 | { | 
|---|
| 7 |     /* This routine returns a uniform pseudorandom double precision number in the | 
|---|
| 8 |        range (0, 1) by using the linear congruential generator | 
|---|
| 9 |  | 
|---|
| 10 |        x_{k+1} = a x_k  (mod 2^46) | 
|---|
| 11 |  | 
|---|
| 12 |        where 0 < x_k < 2^46 and 0 < a < 2^46.  This scheme generates 2^44 numbers | 
|---|
| 13 |        before repeating.  The argument A is the same as 'a' in the above formula, | 
|---|
| 14 |        and X is the same as x_0.  A and X must be odd double precision integers | 
|---|
| 15 |        in the range (1, 2^46).  The returned value RANDLC is normalized to be | 
|---|
| 16 |        between 0 and 1, i.e. RANDLC = 2^(-46) * x_1.  X is updated to contain | 
|---|
| 17 |        the new seed x_1, so that subsequent calls to RANDLC using the same | 
|---|
| 18 |        arguments will generate a continuous sequence. | 
|---|
| 19 |  | 
|---|
| 20 |        This routine should produce the same results on any computer with at least | 
|---|
| 21 |        48 mantissa bits in double precision floating point data.  On 64 bit | 
|---|
| 22 |        systems, double precision should be disabled. | 
|---|
| 23 |  | 
|---|
| 24 |        David H. Bailey     October 26, 1990 */ | 
|---|
| 25 |  | 
|---|
| 26 |     unsigned long long i246m1, Lx, La; | 
|---|
| 27 |     double d2m46; | 
|---|
| 28 |  | 
|---|
| 29 |     d2m46 = 0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5* | 
|---|
| 30 |         0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5* | 
|---|
| 31 |         0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5* | 
|---|
| 32 |         0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5* | 
|---|
| 33 |         0.5*0.5*0.5*0.5*0.5*0.5; | 
|---|
| 34 |     //d2m46 = pow( 0.5, 46 ); | 
|---|
| 35 |  | 
|---|
| 36 |     i246m1 = 0x00003FFFFFFFFFFFLL; | 
|---|
| 37 |  | 
|---|
| 38 |     Lx = *x; | 
|---|
| 39 |     La = a; | 
|---|
| 40 |  | 
|---|
| 41 |     Lx = (Lx*La)&i246m1; | 
|---|
| 42 |     *x = (double) Lx; | 
|---|
| 43 |     return (d2m46 * (*x)); | 
|---|
| 44 | } | 
|---|
| 45 |  | 
|---|
| 46 | void vranlc (int n, double *x_seed, double a, double y[]) { | 
|---|
| 47 |     /* This routine generates N uniform pseudorandom double precision numbers in | 
|---|
| 48 |        the range (0, 1) by using the linear congruential generator | 
|---|
| 49 |  | 
|---|
| 50 |        x_{k+1} = a x_k  (mod 2^46) | 
|---|
| 51 |  | 
|---|
| 52 |        where 0 < x_k < 2^46 and 0 < a < 2^46.  This scheme generates 2^44 numbers | 
|---|
| 53 |        before repeating.  The argument A is the same as 'a' in the above formula, | 
|---|
| 54 |        and X is the same as x_0.  A and X must be odd double precision integers | 
|---|
| 55 |        in the range (1, 2^46).  The N results are placed in Y and are normalized | 
|---|
| 56 |        to be between 0 and 1.  X is updated to contain the new seed, so that | 
|---|
| 57 |        subsequent calls to VRANLC using the same arguments will generate a | 
|---|
| 58 |        continuous sequence.  If N is zero, only initialization is performed, and | 
|---|
| 59 |        the variables X, A and Y are ignored. | 
|---|
| 60 |  | 
|---|
| 61 |        This routine is the standard version designed for scalar or RISC systems. | 
|---|
| 62 |        However, it should produce the same results on any single processor | 
|---|
| 63 |        computer with at least 48 mantissa bits in double precision floating point | 
|---|
| 64 |        data.  On 64 bit systems, double precision should be disabled. | 
|---|
| 65 |      */ | 
|---|
| 66 |     int i; | 
|---|
| 67 |     double x; | 
|---|
| 68 |     unsigned long long i246m1, Lx, La; | 
|---|
| 69 |     double d2m46; | 
|---|
| 70 |  | 
|---|
| 71 |     d2m46 = 0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5* | 
|---|
| 72 |         0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5* | 
|---|
| 73 |         0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5* | 
|---|
| 74 |         0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5*0.5* | 
|---|
| 75 |         0.5*0.5*0.5*0.5*0.5*0.5; | 
|---|
| 76 |     //  d2m46 = pow( 0.5, 46.0 ); | 
|---|
| 77 |     i246m1 = 0x00003FFFFFFFFFFFLL; | 
|---|
| 78 |  | 
|---|
| 79 |     x = *x_seed; | 
|---|
| 80 |     Lx = x; | 
|---|
| 81 |     La = a; | 
|---|
| 82 |  | 
|---|
| 83 |     for (i = 1; i <= n; i++) | 
|---|
| 84 |     { | 
|---|
| 85 |         Lx = ((Lx*La)&i246m1); | 
|---|
| 86 |         x = (double) Lx; | 
|---|
| 87 |         y[i] = d2m46 * x; | 
|---|
| 88 |     } | 
|---|
| 89 |     *x_seed = x; | 
|---|
| 90 | } | 
|---|