1 SUBROUTINE DCHKGL( NIN, NOUT )
2 *
3 * -- LAPACK test routine (version 3.1) --
4 * Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
5 * November 2006
6 *
7 * .. Scalar Arguments ..
8 INTEGER NIN, NOUT
9 * ..
10 *
11 * Purpose
12 * =======
13 *
14 * DCHKGL tests DGGBAL, a routine for balancing a matrix pair (A, B).
15 *
16 * Arguments
17 * =========
18 *
19 * NIN (input) INTEGER
20 * The logical unit number for input. NIN > 0.
21 *
22 * NOUT (input) INTEGER
23 * The logical unit number for output. NOUT > 0.
24 *
25 * =====================================================================
26 *
27 * .. Parameters ..
28 INTEGER LDA, LDB, LWORK
29 PARAMETER ( LDA = 20, LDB = 20, LWORK = 6*LDA )
30 DOUBLE PRECISION ZERO
31 PARAMETER ( ZERO = 0.0D+0 )
32 * ..
33 * .. Local Scalars ..
34 INTEGER I, IHI, IHIIN, ILO, ILOIN, INFO, J, KNT, N,
35 $ NINFO
36 DOUBLE PRECISION ANORM, BNORM, EPS, RMAX, VMAX
37 * ..
38 * .. Local Arrays ..
39 INTEGER LMAX( 5 )
40 DOUBLE PRECISION A( LDA, LDA ), AIN( LDA, LDA ), B( LDB, LDB ),
41 $ BIN( LDB, LDB ), LSCALE( LDA ), LSCLIN( LDA ),
42 $ RSCALE( LDA ), RSCLIN( LDA ), WORK( LWORK )
43 * ..
44 * .. External Functions ..
45 DOUBLE PRECISION DLAMCH, DLANGE
46 EXTERNAL DLAMCH, DLANGE
47 * ..
48 * .. External Subroutines ..
49 EXTERNAL DGGBAL
50 * ..
51 * .. Intrinsic Functions ..
52 INTRINSIC ABS, MAX
53 * ..
54 * .. Executable Statements ..
55 *
56 LMAX( 1 ) = 0
57 LMAX( 2 ) = 0
58 LMAX( 3 ) = 0
59 NINFO = 0
60 KNT = 0
61 RMAX = ZERO
62 *
63 EPS = DLAMCH( 'Precision' )
64 *
65 10 CONTINUE
66 *
67 READ( NIN, FMT = * )N
68 IF( N.EQ.0 )
69 $ GO TO 90
70 DO 20 I = 1, N
71 READ( NIN, FMT = * )( A( I, J ), J = 1, N )
72 20 CONTINUE
73 *
74 DO 30 I = 1, N
75 READ( NIN, FMT = * )( B( I, J ), J = 1, N )
76 30 CONTINUE
77 *
78 READ( NIN, FMT = * )ILOIN, IHIIN
79 DO 40 I = 1, N
80 READ( NIN, FMT = * )( AIN( I, J ), J = 1, N )
81 40 CONTINUE
82 DO 50 I = 1, N
83 READ( NIN, FMT = * )( BIN( I, J ), J = 1, N )
84 50 CONTINUE
85 *
86 READ( NIN, FMT = * )( LSCLIN( I ), I = 1, N )
87 READ( NIN, FMT = * )( RSCLIN( I ), I = 1, N )
88 *
89 ANORM = DLANGE( 'M', N, N, A, LDA, WORK )
90 BNORM = DLANGE( 'M', N, N, B, LDB, WORK )
91 *
92 KNT = KNT + 1
93 *
94 CALL DGGBAL( 'B', N, A, LDA, B, LDB, ILO, IHI, LSCALE, RSCALE,
95 $ WORK, INFO )
96 *
97 IF( INFO.NE.0 ) THEN
98 NINFO = NINFO + 1
99 LMAX( 1 ) = KNT
100 END IF
101 *
102 IF( ILO.NE.ILOIN .OR. IHI.NE.IHIIN ) THEN
103 NINFO = NINFO + 1
104 LMAX( 2 ) = KNT
105 END IF
106 *
107 VMAX = ZERO
108 DO 70 I = 1, N
109 DO 60 J = 1, N
110 VMAX = MAX( VMAX, ABS( A( I, J )-AIN( I, J ) ) )
111 VMAX = MAX( VMAX, ABS( B( I, J )-BIN( I, J ) ) )
112 60 CONTINUE
113 70 CONTINUE
114 *
115 DO 80 I = 1, N
116 VMAX = MAX( VMAX, ABS( LSCALE( I )-LSCLIN( I ) ) )
117 VMAX = MAX( VMAX, ABS( RSCALE( I )-RSCLIN( I ) ) )
118 80 CONTINUE
119 *
120 VMAX = VMAX / ( EPS*MAX( ANORM, BNORM ) )
121 *
122 IF( VMAX.GT.RMAX ) THEN
123 LMAX( 3 ) = KNT
124 RMAX = VMAX
125 END IF
126 *
127 GO TO 10
128 *
129 90 CONTINUE
130 *
131 WRITE( NOUT, FMT = 9999 )
132 9999 FORMAT( 1X, '.. test output of DGGBAL .. ' )
133 *
134 WRITE( NOUT, FMT = 9998 )RMAX
135 9998 FORMAT( 1X, 'value of largest test error = ', D12.3 )
136 WRITE( NOUT, FMT = 9997 )LMAX( 1 )
137 9997 FORMAT( 1X, 'example number where info is not zero = ', I4 )
138 WRITE( NOUT, FMT = 9996 )LMAX( 2 )
139 9996 FORMAT( 1X, 'example number where ILO or IHI wrong = ', I4 )
140 WRITE( NOUT, FMT = 9995 )LMAX( 3 )
141 9995 FORMAT( 1X, 'example number having largest error = ', I4 )
142 WRITE( NOUT, FMT = 9994 )NINFO
143 9994 FORMAT( 1X, 'number of examples where info is not 0 = ', I4 )
144 WRITE( NOUT, FMT = 9993 )KNT
145 9993 FORMAT( 1X, 'total number of examples tested = ', I4 )
146 *
147 RETURN
148 *
149 * End of DCHKGL
150 *
151 END
2 *
3 * -- LAPACK test routine (version 3.1) --
4 * Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
5 * November 2006
6 *
7 * .. Scalar Arguments ..
8 INTEGER NIN, NOUT
9 * ..
10 *
11 * Purpose
12 * =======
13 *
14 * DCHKGL tests DGGBAL, a routine for balancing a matrix pair (A, B).
15 *
16 * Arguments
17 * =========
18 *
19 * NIN (input) INTEGER
20 * The logical unit number for input. NIN > 0.
21 *
22 * NOUT (input) INTEGER
23 * The logical unit number for output. NOUT > 0.
24 *
25 * =====================================================================
26 *
27 * .. Parameters ..
28 INTEGER LDA, LDB, LWORK
29 PARAMETER ( LDA = 20, LDB = 20, LWORK = 6*LDA )
30 DOUBLE PRECISION ZERO
31 PARAMETER ( ZERO = 0.0D+0 )
32 * ..
33 * .. Local Scalars ..
34 INTEGER I, IHI, IHIIN, ILO, ILOIN, INFO, J, KNT, N,
35 $ NINFO
36 DOUBLE PRECISION ANORM, BNORM, EPS, RMAX, VMAX
37 * ..
38 * .. Local Arrays ..
39 INTEGER LMAX( 5 )
40 DOUBLE PRECISION A( LDA, LDA ), AIN( LDA, LDA ), B( LDB, LDB ),
41 $ BIN( LDB, LDB ), LSCALE( LDA ), LSCLIN( LDA ),
42 $ RSCALE( LDA ), RSCLIN( LDA ), WORK( LWORK )
43 * ..
44 * .. External Functions ..
45 DOUBLE PRECISION DLAMCH, DLANGE
46 EXTERNAL DLAMCH, DLANGE
47 * ..
48 * .. External Subroutines ..
49 EXTERNAL DGGBAL
50 * ..
51 * .. Intrinsic Functions ..
52 INTRINSIC ABS, MAX
53 * ..
54 * .. Executable Statements ..
55 *
56 LMAX( 1 ) = 0
57 LMAX( 2 ) = 0
58 LMAX( 3 ) = 0
59 NINFO = 0
60 KNT = 0
61 RMAX = ZERO
62 *
63 EPS = DLAMCH( 'Precision' )
64 *
65 10 CONTINUE
66 *
67 READ( NIN, FMT = * )N
68 IF( N.EQ.0 )
69 $ GO TO 90
70 DO 20 I = 1, N
71 READ( NIN, FMT = * )( A( I, J ), J = 1, N )
72 20 CONTINUE
73 *
74 DO 30 I = 1, N
75 READ( NIN, FMT = * )( B( I, J ), J = 1, N )
76 30 CONTINUE
77 *
78 READ( NIN, FMT = * )ILOIN, IHIIN
79 DO 40 I = 1, N
80 READ( NIN, FMT = * )( AIN( I, J ), J = 1, N )
81 40 CONTINUE
82 DO 50 I = 1, N
83 READ( NIN, FMT = * )( BIN( I, J ), J = 1, N )
84 50 CONTINUE
85 *
86 READ( NIN, FMT = * )( LSCLIN( I ), I = 1, N )
87 READ( NIN, FMT = * )( RSCLIN( I ), I = 1, N )
88 *
89 ANORM = DLANGE( 'M', N, N, A, LDA, WORK )
90 BNORM = DLANGE( 'M', N, N, B, LDB, WORK )
91 *
92 KNT = KNT + 1
93 *
94 CALL DGGBAL( 'B', N, A, LDA, B, LDB, ILO, IHI, LSCALE, RSCALE,
95 $ WORK, INFO )
96 *
97 IF( INFO.NE.0 ) THEN
98 NINFO = NINFO + 1
99 LMAX( 1 ) = KNT
100 END IF
101 *
102 IF( ILO.NE.ILOIN .OR. IHI.NE.IHIIN ) THEN
103 NINFO = NINFO + 1
104 LMAX( 2 ) = KNT
105 END IF
106 *
107 VMAX = ZERO
108 DO 70 I = 1, N
109 DO 60 J = 1, N
110 VMAX = MAX( VMAX, ABS( A( I, J )-AIN( I, J ) ) )
111 VMAX = MAX( VMAX, ABS( B( I, J )-BIN( I, J ) ) )
112 60 CONTINUE
113 70 CONTINUE
114 *
115 DO 80 I = 1, N
116 VMAX = MAX( VMAX, ABS( LSCALE( I )-LSCLIN( I ) ) )
117 VMAX = MAX( VMAX, ABS( RSCALE( I )-RSCLIN( I ) ) )
118 80 CONTINUE
119 *
120 VMAX = VMAX / ( EPS*MAX( ANORM, BNORM ) )
121 *
122 IF( VMAX.GT.RMAX ) THEN
123 LMAX( 3 ) = KNT
124 RMAX = VMAX
125 END IF
126 *
127 GO TO 10
128 *
129 90 CONTINUE
130 *
131 WRITE( NOUT, FMT = 9999 )
132 9999 FORMAT( 1X, '.. test output of DGGBAL .. ' )
133 *
134 WRITE( NOUT, FMT = 9998 )RMAX
135 9998 FORMAT( 1X, 'value of largest test error = ', D12.3 )
136 WRITE( NOUT, FMT = 9997 )LMAX( 1 )
137 9997 FORMAT( 1X, 'example number where info is not zero = ', I4 )
138 WRITE( NOUT, FMT = 9996 )LMAX( 2 )
139 9996 FORMAT( 1X, 'example number where ILO or IHI wrong = ', I4 )
140 WRITE( NOUT, FMT = 9995 )LMAX( 3 )
141 9995 FORMAT( 1X, 'example number having largest error = ', I4 )
142 WRITE( NOUT, FMT = 9994 )NINFO
143 9994 FORMAT( 1X, 'number of examples where info is not 0 = ', I4 )
144 WRITE( NOUT, FMT = 9993 )KNT
145 9993 FORMAT( 1X, 'total number of examples tested = ', I4 )
146 *
147 RETURN
148 *
149 * End of DCHKGL
150 *
151 END