1: *> \brief \b DGBEQU
2: *
3: * =========== DOCUMENTATION ===========
4: *
5: * Online html documentation available at
6: * http://www.netlib.org/lapack/explore-html/
7: *
8: *> \htmlonly
9: *> Download DGBEQU + dependencies
10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dgbequ.f">
11: *> [TGZ]</a>
12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dgbequ.f">
13: *> [ZIP]</a>
14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dgbequ.f">
15: *> [TXT]</a>
16: *> \endhtmlonly
17: *
18: * Definition:
19: * ===========
20: *
21: * SUBROUTINE DGBEQU( M, N, KL, KU, AB, LDAB, R, C, ROWCND, COLCND,
22: * AMAX, INFO )
23: *
24: * .. Scalar Arguments ..
25: * INTEGER INFO, KL, KU, LDAB, M, N
26: * DOUBLE PRECISION AMAX, COLCND, ROWCND
27: * ..
28: * .. Array Arguments ..
29: * DOUBLE PRECISION AB( LDAB, * ), C( * ), R( * )
30: * ..
31: *
32: *
33: *> \par Purpose:
34: * =============
35: *>
36: *> \verbatim
37: *>
38: *> DGBEQU computes row and column scalings intended to equilibrate an
39: *> M-by-N band matrix A and reduce its condition number. R returns the
40: *> row scale factors and C the column scale factors, chosen to try to
41: *> make the largest element in each row and column of the matrix B with
42: *> elements B(i,j)=R(i)*A(i,j)*C(j) have absolute value 1.
43: *>
44: *> R(i) and C(j) are restricted to be between SMLNUM = smallest safe
45: *> number and BIGNUM = largest safe number. Use of these scaling
46: *> factors is not guaranteed to reduce the condition number of A but
47: *> works well in practice.
48: *> \endverbatim
49: *
50: * Arguments:
51: * ==========
52: *
53: *> \param[in] M
54: *> \verbatim
55: *> M is INTEGER
56: *> The number of rows of the matrix A. M >= 0.
57: *> \endverbatim
58: *>
59: *> \param[in] N
60: *> \verbatim
61: *> N is INTEGER
62: *> The number of columns of the matrix A. N >= 0.
63: *> \endverbatim
64: *>
65: *> \param[in] KL
66: *> \verbatim
67: *> KL is INTEGER
68: *> The number of subdiagonals within the band of A. KL >= 0.
69: *> \endverbatim
70: *>
71: *> \param[in] KU
72: *> \verbatim
73: *> KU is INTEGER
74: *> The number of superdiagonals within the band of A. KU >= 0.
75: *> \endverbatim
76: *>
77: *> \param[in] AB
78: *> \verbatim
79: *> AB is DOUBLE PRECISION array, dimension (LDAB,N)
80: *> The band matrix A, stored in rows 1 to KL+KU+1. The j-th
81: *> column of A is stored in the j-th column of the array AB as
82: *> follows:
83: *> AB(ku+1+i-j,j) = A(i,j) for max(1,j-ku)<=i<=min(m,j+kl).
84: *> \endverbatim
85: *>
86: *> \param[in] LDAB
87: *> \verbatim
88: *> LDAB is INTEGER
89: *> The leading dimension of the array AB. LDAB >= KL+KU+1.
90: *> \endverbatim
91: *>
92: *> \param[out] R
93: *> \verbatim
94: *> R is DOUBLE PRECISION array, dimension (M)
95: *> If INFO = 0, or INFO > M, R contains the row scale factors
96: *> for A.
97: *> \endverbatim
98: *>
99: *> \param[out] C
100: *> \verbatim
101: *> C is DOUBLE PRECISION array, dimension (N)
102: *> If INFO = 0, C contains the column scale factors for A.
103: *> \endverbatim
104: *>
105: *> \param[out] ROWCND
106: *> \verbatim
107: *> ROWCND is DOUBLE PRECISION
108: *> If INFO = 0 or INFO > M, ROWCND contains the ratio of the
109: *> smallest R(i) to the largest R(i). If ROWCND >= 0.1 and
110: *> AMAX is neither too large nor too small, it is not worth
111: *> scaling by R.
112: *> \endverbatim
113: *>
114: *> \param[out] COLCND
115: *> \verbatim
116: *> COLCND is DOUBLE PRECISION
117: *> If INFO = 0, COLCND contains the ratio of the smallest
118: *> C(i) to the largest C(i). If COLCND >= 0.1, it is not
119: *> worth scaling by C.
120: *> \endverbatim
121: *>
122: *> \param[out] AMAX
123: *> \verbatim
124: *> AMAX is DOUBLE PRECISION
125: *> Absolute value of largest matrix element. If AMAX is very
126: *> close to overflow or very close to underflow, the matrix
127: *> should be scaled.
128: *> \endverbatim
129: *>
130: *> \param[out] INFO
131: *> \verbatim
132: *> INFO is INTEGER
133: *> = 0: successful exit
134: *> < 0: if INFO = -i, the i-th argument had an illegal value
135: *> > 0: if INFO = i, and i is
136: *> <= M: the i-th row of A is exactly zero
137: *> > M: the (i-M)-th column of A is exactly zero
138: *> \endverbatim
139: *
140: * Authors:
141: * ========
142: *
143: *> \author Univ. of Tennessee
144: *> \author Univ. of California Berkeley
145: *> \author Univ. of Colorado Denver
146: *> \author NAG Ltd.
147: *
148: *> \date November 2011
149: *
150: *> \ingroup doubleGBcomputational
151: *
152: * =====================================================================
153: SUBROUTINE DGBEQU( M, N, KL, KU, AB, LDAB, R, C, ROWCND, COLCND,
154: $ AMAX, INFO )
155: *
156: * -- LAPACK computational routine (version 3.4.0) --
157: * -- LAPACK is a software package provided by Univ. of Tennessee, --
158: * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
159: * November 2011
160: *
161: * .. Scalar Arguments ..
162: INTEGER INFO, KL, KU, LDAB, M, N
163: DOUBLE PRECISION AMAX, COLCND, ROWCND
164: * ..
165: * .. Array Arguments ..
166: DOUBLE PRECISION AB( LDAB, * ), C( * ), R( * )
167: * ..
168: *
169: * =====================================================================
170: *
171: * .. Parameters ..
172: DOUBLE PRECISION ONE, ZERO
173: PARAMETER ( ONE = 1.0D+0, ZERO = 0.0D+0 )
174: * ..
175: * .. Local Scalars ..
176: INTEGER I, J, KD
177: DOUBLE PRECISION BIGNUM, RCMAX, RCMIN, SMLNUM
178: * ..
179: * .. External Functions ..
180: DOUBLE PRECISION DLAMCH
181: EXTERNAL DLAMCH
182: * ..
183: * .. External Subroutines ..
184: EXTERNAL XERBLA
185: * ..
186: * .. Intrinsic Functions ..
187: INTRINSIC ABS, MAX, MIN
188: * ..
189: * .. Executable Statements ..
190: *
191: * Test the input parameters
192: *
193: INFO = 0
194: IF( M.LT.0 ) THEN
195: INFO = -1
196: ELSE IF( N.LT.0 ) THEN
197: INFO = -2
198: ELSE IF( KL.LT.0 ) THEN
199: INFO = -3
200: ELSE IF( KU.LT.0 ) THEN
201: INFO = -4
202: ELSE IF( LDAB.LT.KL+KU+1 ) THEN
203: INFO = -6
204: END IF
205: IF( INFO.NE.0 ) THEN
206: CALL XERBLA( 'DGBEQU', -INFO )
207: RETURN
208: END IF
209: *
210: * Quick return if possible
211: *
212: IF( M.EQ.0 .OR. N.EQ.0 ) THEN
213: ROWCND = ONE
214: COLCND = ONE
215: AMAX = ZERO
216: RETURN
217: END IF
218: *
219: * Get machine constants.
220: *
221: SMLNUM = DLAMCH( 'S' )
222: BIGNUM = ONE / SMLNUM
223: *
224: * Compute row scale factors.
225: *
226: DO 10 I = 1, M
227: R( I ) = ZERO
228: 10 CONTINUE
229: *
230: * Find the maximum element in each row.
231: *
232: KD = KU + 1
233: DO 30 J = 1, N
234: DO 20 I = MAX( J-KU, 1 ), MIN( J+KL, M )
235: R( I ) = MAX( R( I ), ABS( AB( KD+I-J, J ) ) )
236: 20 CONTINUE
237: 30 CONTINUE
238: *
239: * Find the maximum and minimum scale factors.
240: *
241: RCMIN = BIGNUM
242: RCMAX = ZERO
243: DO 40 I = 1, M
244: RCMAX = MAX( RCMAX, R( I ) )
245: RCMIN = MIN( RCMIN, R( I ) )
246: 40 CONTINUE
247: AMAX = RCMAX
248: *
249: IF( RCMIN.EQ.ZERO ) THEN
250: *
251: * Find the first zero scale factor and return an error code.
252: *
253: DO 50 I = 1, M
254: IF( R( I ).EQ.ZERO ) THEN
255: INFO = I
256: RETURN
257: END IF
258: 50 CONTINUE
259: ELSE
260: *
261: * Invert the scale factors.
262: *
263: DO 60 I = 1, M
264: R( I ) = ONE / MIN( MAX( R( I ), SMLNUM ), BIGNUM )
265: 60 CONTINUE
266: *
267: * Compute ROWCND = min(R(I)) / max(R(I))
268: *
269: ROWCND = MAX( RCMIN, SMLNUM ) / MIN( RCMAX, BIGNUM )
270: END IF
271: *
272: * Compute column scale factors
273: *
274: DO 70 J = 1, N
275: C( J ) = ZERO
276: 70 CONTINUE
277: *
278: * Find the maximum element in each column,
279: * assuming the row scaling computed above.
280: *
281: KD = KU + 1
282: DO 90 J = 1, N
283: DO 80 I = MAX( J-KU, 1 ), MIN( J+KL, M )
284: C( J ) = MAX( C( J ), ABS( AB( KD+I-J, J ) )*R( I ) )
285: 80 CONTINUE
286: 90 CONTINUE
287: *
288: * Find the maximum and minimum scale factors.
289: *
290: RCMIN = BIGNUM
291: RCMAX = ZERO
292: DO 100 J = 1, N
293: RCMIN = MIN( RCMIN, C( J ) )
294: RCMAX = MAX( RCMAX, C( J ) )
295: 100 CONTINUE
296: *
297: IF( RCMIN.EQ.ZERO ) THEN
298: *
299: * Find the first zero scale factor and return an error code.
300: *
301: DO 110 J = 1, N
302: IF( C( J ).EQ.ZERO ) THEN
303: INFO = M + J
304: RETURN
305: END IF
306: 110 CONTINUE
307: ELSE
308: *
309: * Invert the scale factors.
310: *
311: DO 120 J = 1, N
312: C( J ) = ONE / MIN( MAX( C( J ), SMLNUM ), BIGNUM )
313: 120 CONTINUE
314: *
315: * Compute COLCND = min(C(J)) / max(C(J))
316: *
317: COLCND = MAX( RCMIN, SMLNUM ) / MIN( RCMAX, BIGNUM )
318: END IF
319: *
320: RETURN
321: *
322: * End of DGBEQU
323: *
324: END
CVSweb interface <joel.bertrand@systella.fr>