00001 SUBROUTINE DLASWP( N, A, LDA, K1, K2, IPIV, INCX ) 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 INCX, K1, K2, LDA, N 00010 * .. 00011 * .. Array Arguments .. 00012 INTEGER IPIV( * ) 00013 DOUBLE PRECISION A( LDA, * ) 00014 * .. 00015 * 00016 * Purpose 00017 * ======= 00018 * 00019 * DLASWP performs a series of row interchanges on the matrix A. 00020 * One row interchange is initiated for each of rows K1 through K2 of A. 00021 * 00022 * Arguments 00023 * ========= 00024 * 00025 * N (input) INTEGER 00026 * The number of columns of the matrix A. 00027 * 00028 * A (input/output) DOUBLE PRECISION array, dimension (LDA,N) 00029 * On entry, the matrix of column dimension N to which the row 00030 * interchanges will be applied. 00031 * On exit, the permuted matrix. 00032 * 00033 * LDA (input) INTEGER 00034 * The leading dimension of the array A. 00035 * 00036 * K1 (input) INTEGER 00037 * The first element of IPIV for which a row interchange will 00038 * be done. 00039 * 00040 * K2 (input) INTEGER 00041 * The last element of IPIV for which a row interchange will 00042 * be done. 00043 * 00044 * IPIV (input) INTEGER array, dimension (K2*abs(INCX)) 00045 * The vector of pivot indices. Only the elements in positions 00046 * K1 through K2 of IPIV are accessed. 00047 * IPIV(K) = L implies rows K and L are to be interchanged. 00048 * 00049 * INCX (input) INTEGER 00050 * The increment between successive values of IPIV. If IPIV 00051 * is negative, the pivots are applied in reverse order. 00052 * 00053 * Further Details 00054 * =============== 00055 * 00056 * Modified by 00057 * R. C. Whaley, Computer Science Dept., Univ. of Tenn., Knoxville, USA 00058 * 00059 * ===================================================================== 00060 * 00061 * .. Local Scalars .. 00062 INTEGER I, I1, I2, INC, IP, IX, IX0, J, K, N32 00063 DOUBLE PRECISION TEMP 00064 * .. 00065 * .. Executable Statements .. 00066 * 00067 * Interchange row I with row IPIV(I) for each of rows K1 through K2. 00068 * 00069 IF( INCX.GT.0 ) THEN 00070 IX0 = K1 00071 I1 = K1 00072 I2 = K2 00073 INC = 1 00074 ELSE IF( INCX.LT.0 ) THEN 00075 IX0 = 1 + ( 1-K2 )*INCX 00076 I1 = K2 00077 I2 = K1 00078 INC = -1 00079 ELSE 00080 RETURN 00081 END IF 00082 * 00083 N32 = ( N / 32 )*32 00084 IF( N32.NE.0 ) THEN 00085 DO 30 J = 1, N32, 32 00086 IX = IX0 00087 DO 20 I = I1, I2, INC 00088 IP = IPIV( IX ) 00089 IF( IP.NE.I ) THEN 00090 DO 10 K = J, J + 31 00091 TEMP = A( I, K ) 00092 A( I, K ) = A( IP, K ) 00093 A( IP, K ) = TEMP 00094 10 CONTINUE 00095 END IF 00096 IX = IX + INCX 00097 20 CONTINUE 00098 30 CONTINUE 00099 END IF 00100 IF( N32.NE.N ) THEN 00101 N32 = N32 + 1 00102 IX = IX0 00103 DO 50 I = I1, I2, INC 00104 IP = IPIV( IX ) 00105 IF( IP.NE.I ) THEN 00106 DO 40 K = N32, N 00107 TEMP = A( I, K ) 00108 A( I, K ) = A( IP, K ) 00109 A( IP, K ) = TEMP 00110 40 CONTINUE 00111 END IF 00112 IX = IX + INCX 00113 50 CONTINUE 00114 END IF 00115 * 00116 RETURN 00117 * 00118 * End of DLASWP 00119 * 00120 END