LAPACK 3.3.1
Linear Algebra PACKage
|
00001 SUBROUTINE ZLARNV( IDIST, ISEED, N, X ) 00002 * 00003 * -- LAPACK auxiliary routine (version 3.2) -- 00004 * -- LAPACK is a software package provided by Univ. of Tennessee, -- 00005 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..-- 00006 * November 2006 00007 * 00008 * .. Scalar Arguments .. 00009 INTEGER IDIST, N 00010 * .. 00011 * .. Array Arguments .. 00012 INTEGER ISEED( 4 ) 00013 COMPLEX*16 X( * ) 00014 * .. 00015 * 00016 * Purpose 00017 * ======= 00018 * 00019 * ZLARNV returns a vector of n random complex numbers from a uniform or 00020 * normal distribution. 00021 * 00022 * Arguments 00023 * ========= 00024 * 00025 * IDIST (input) INTEGER 00026 * Specifies the distribution of the random numbers: 00027 * = 1: real and imaginary parts each uniform (0,1) 00028 * = 2: real and imaginary parts each uniform (-1,1) 00029 * = 3: real and imaginary parts each normal (0,1) 00030 * = 4: uniformly distributed on the disc abs(z) < 1 00031 * = 5: uniformly distributed on the circle abs(z) = 1 00032 * 00033 * ISEED (input/output) INTEGER array, dimension (4) 00034 * On entry, the seed of the random number generator; the array 00035 * elements must be between 0 and 4095, and ISEED(4) must be 00036 * odd. 00037 * On exit, the seed is updated. 00038 * 00039 * N (input) INTEGER 00040 * The number of random numbers to be generated. 00041 * 00042 * X (output) COMPLEX*16 array, dimension (N) 00043 * The generated random numbers. 00044 * 00045 * Further Details 00046 * =============== 00047 * 00048 * This routine calls the auxiliary routine DLARUV to generate random 00049 * real numbers from a uniform (0,1) distribution, in batches of up to 00050 * 128 using vectorisable code. The Box-Muller method is used to 00051 * transform numbers from a uniform to a normal distribution. 00052 * 00053 * ===================================================================== 00054 * 00055 * .. Parameters .. 00056 DOUBLE PRECISION ZERO, ONE, TWO 00057 PARAMETER ( ZERO = 0.0D+0, ONE = 1.0D+0, TWO = 2.0D+0 ) 00058 INTEGER LV 00059 PARAMETER ( LV = 128 ) 00060 DOUBLE PRECISION TWOPI 00061 PARAMETER ( TWOPI = 6.2831853071795864769252867663D+0 ) 00062 * .. 00063 * .. Local Scalars .. 00064 INTEGER I, IL, IV 00065 * .. 00066 * .. Local Arrays .. 00067 DOUBLE PRECISION U( LV ) 00068 * .. 00069 * .. Intrinsic Functions .. 00070 INTRINSIC DCMPLX, EXP, LOG, MIN, SQRT 00071 * .. 00072 * .. External Subroutines .. 00073 EXTERNAL DLARUV 00074 * .. 00075 * .. Executable Statements .. 00076 * 00077 DO 60 IV = 1, N, LV / 2 00078 IL = MIN( LV / 2, N-IV+1 ) 00079 * 00080 * Call DLARUV to generate 2*IL real numbers from a uniform (0,1) 00081 * distribution (2*IL <= LV) 00082 * 00083 CALL DLARUV( ISEED, 2*IL, U ) 00084 * 00085 IF( IDIST.EQ.1 ) THEN 00086 * 00087 * Copy generated numbers 00088 * 00089 DO 10 I = 1, IL 00090 X( IV+I-1 ) = DCMPLX( U( 2*I-1 ), U( 2*I ) ) 00091 10 CONTINUE 00092 ELSE IF( IDIST.EQ.2 ) THEN 00093 * 00094 * Convert generated numbers to uniform (-1,1) distribution 00095 * 00096 DO 20 I = 1, IL 00097 X( IV+I-1 ) = DCMPLX( TWO*U( 2*I-1 )-ONE, 00098 $ TWO*U( 2*I )-ONE ) 00099 20 CONTINUE 00100 ELSE IF( IDIST.EQ.3 ) THEN 00101 * 00102 * Convert generated numbers to normal (0,1) distribution 00103 * 00104 DO 30 I = 1, IL 00105 X( IV+I-1 ) = SQRT( -TWO*LOG( U( 2*I-1 ) ) )* 00106 $ EXP( DCMPLX( ZERO, TWOPI*U( 2*I ) ) ) 00107 30 CONTINUE 00108 ELSE IF( IDIST.EQ.4 ) THEN 00109 * 00110 * Convert generated numbers to complex numbers uniformly 00111 * distributed on the unit disk 00112 * 00113 DO 40 I = 1, IL 00114 X( IV+I-1 ) = SQRT( U( 2*I-1 ) )* 00115 $ EXP( DCMPLX( ZERO, TWOPI*U( 2*I ) ) ) 00116 40 CONTINUE 00117 ELSE IF( IDIST.EQ.5 ) THEN 00118 * 00119 * Convert generated numbers to complex numbers uniformly 00120 * distributed on the unit circle 00121 * 00122 DO 50 I = 1, IL 00123 X( IV+I-1 ) = EXP( DCMPLX( ZERO, TWOPI*U( 2*I ) ) ) 00124 50 CONTINUE 00125 END IF 00126 60 CONTINUE 00127 RETURN 00128 * 00129 * End of ZLARNV 00130 * 00131 END