• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* dsbmv.f -- translated by f2c (version 20100827).
2    You must link the resulting object file with libf2c:
3 	on Microsoft Windows system, link with libf2c.lib;
4 	on Linux or Unix systems, link with .../path/to/libf2c.a -lm
5 	or, if you install libf2c.a in a standard place, with -lf2c -lm
6 	-- in that order, at the end of the command line, as in
7 		cc *.o -lf2c -lm
8 	Source for libf2c is in /netlib/f2c/libf2c.zip, e.g.,
9 
10 		http://www.netlib.org/f2c/libf2c.zip
11 */
12 
13 #include "datatypes.h"
14 
dsbmv_(char * uplo,integer * n,integer * k,doublereal * alpha,doublereal * a,integer * lda,doublereal * x,integer * incx,doublereal * beta,doublereal * y,integer * incy,ftnlen uplo_len)15 /* Subroutine */ int dsbmv_(char *uplo, integer *n, integer *k, doublereal *
16 	alpha, doublereal *a, integer *lda, doublereal *x, integer *incx,
17 	doublereal *beta, doublereal *y, integer *incy, ftnlen uplo_len)
18 {
19     /* System generated locals */
20     integer a_dim1, a_offset, i__1, i__2, i__3, i__4;
21 
22     /* Local variables */
23     integer i__, j, l, ix, iy, jx, jy, kx, ky, info;
24     doublereal temp1, temp2;
25     extern logical lsame_(char *, char *, ftnlen, ftnlen);
26     integer kplus1;
27     extern /* Subroutine */ int xerbla_(char *, integer *, ftnlen);
28 
29 /*     .. Scalar Arguments .. */
30 /*     .. */
31 /*     .. Array Arguments .. */
32 /*     .. */
33 
34 /*  Purpose */
35 /*  ======= */
36 
37 /*  DSBMV  performs the matrix-vector  operation */
38 
39 /*     y := alpha*A*x + beta*y, */
40 
41 /*  where alpha and beta are scalars, x and y are n element vectors and */
42 /*  A is an n by n symmetric band matrix, with k super-diagonals. */
43 
44 /*  Arguments */
45 /*  ========== */
46 
47 /*  UPLO   - CHARACTER*1. */
48 /*           On entry, UPLO specifies whether the upper or lower */
49 /*           triangular part of the band matrix A is being supplied as */
50 /*           follows: */
51 
52 /*              UPLO = 'U' or 'u'   The upper triangular part of A is */
53 /*                                  being supplied. */
54 
55 /*              UPLO = 'L' or 'l'   The lower triangular part of A is */
56 /*                                  being supplied. */
57 
58 /*           Unchanged on exit. */
59 
60 /*  N      - INTEGER. */
61 /*           On entry, N specifies the order of the matrix A. */
62 /*           N must be at least zero. */
63 /*           Unchanged on exit. */
64 
65 /*  K      - INTEGER. */
66 /*           On entry, K specifies the number of super-diagonals of the */
67 /*           matrix A. K must satisfy  0 .le. K. */
68 /*           Unchanged on exit. */
69 
70 /*  ALPHA  - DOUBLE PRECISION. */
71 /*           On entry, ALPHA specifies the scalar alpha. */
72 /*           Unchanged on exit. */
73 
74 /*  A      - DOUBLE PRECISION array of DIMENSION ( LDA, n ). */
75 /*           Before entry with UPLO = 'U' or 'u', the leading ( k + 1 ) */
76 /*           by n part of the array A must contain the upper triangular */
77 /*           band part of the symmetric matrix, supplied column by */
78 /*           column, with the leading diagonal of the matrix in row */
79 /*           ( k + 1 ) of the array, the first super-diagonal starting at */
80 /*           position 2 in row k, and so on. The top left k by k triangle */
81 /*           of the array A is not referenced. */
82 /*           The following program segment will transfer the upper */
83 /*           triangular part of a symmetric band matrix from conventional */
84 /*           full matrix storage to band storage: */
85 
86 /*                 DO 20, J = 1, N */
87 /*                    M = K + 1 - J */
88 /*                    DO 10, I = MAX( 1, J - K ), J */
89 /*                       A( M + I, J ) = matrix( I, J ) */
90 /*              10    CONTINUE */
91 /*              20 CONTINUE */
92 
93 /*           Before entry with UPLO = 'L' or 'l', the leading ( k + 1 ) */
94 /*           by n part of the array A must contain the lower triangular */
95 /*           band part of the symmetric matrix, supplied column by */
96 /*           column, with the leading diagonal of the matrix in row 1 of */
97 /*           the array, the first sub-diagonal starting at position 1 in */
98 /*           row 2, and so on. The bottom right k by k triangle of the */
99 /*           array A is not referenced. */
100 /*           The following program segment will transfer the lower */
101 /*           triangular part of a symmetric band matrix from conventional */
102 /*           full matrix storage to band storage: */
103 
104 /*                 DO 20, J = 1, N */
105 /*                    M = 1 - J */
106 /*                    DO 10, I = J, MIN( N, J + K ) */
107 /*                       A( M + I, J ) = matrix( I, J ) */
108 /*              10    CONTINUE */
109 /*              20 CONTINUE */
110 
111 /*           Unchanged on exit. */
112 
113 /*  LDA    - INTEGER. */
114 /*           On entry, LDA specifies the first dimension of A as declared */
115 /*           in the calling (sub) program. LDA must be at least */
116 /*           ( k + 1 ). */
117 /*           Unchanged on exit. */
118 
119 /*  X      - DOUBLE PRECISION array of DIMENSION at least */
120 /*           ( 1 + ( n - 1 )*abs( INCX ) ). */
121 /*           Before entry, the incremented array X must contain the */
122 /*           vector x. */
123 /*           Unchanged on exit. */
124 
125 /*  INCX   - INTEGER. */
126 /*           On entry, INCX specifies the increment for the elements of */
127 /*           X. INCX must not be zero. */
128 /*           Unchanged on exit. */
129 
130 /*  BETA   - DOUBLE PRECISION. */
131 /*           On entry, BETA specifies the scalar beta. */
132 /*           Unchanged on exit. */
133 
134 /*  Y      - DOUBLE PRECISION array of DIMENSION at least */
135 /*           ( 1 + ( n - 1 )*abs( INCY ) ). */
136 /*           Before entry, the incremented array Y must contain the */
137 /*           vector y. On exit, Y is overwritten by the updated vector y. */
138 
139 /*  INCY   - INTEGER. */
140 /*           On entry, INCY specifies the increment for the elements of */
141 /*           Y. INCY must not be zero. */
142 /*           Unchanged on exit. */
143 
144 
145 /*  Level 2 Blas routine. */
146 
147 /*  -- Written on 22-October-1986. */
148 /*     Jack Dongarra, Argonne National Lab. */
149 /*     Jeremy Du Croz, Nag Central Office. */
150 /*     Sven Hammarling, Nag Central Office. */
151 /*     Richard Hanson, Sandia National Labs. */
152 
153 /*  ===================================================================== */
154 
155 /*     .. Parameters .. */
156 /*     .. */
157 /*     .. Local Scalars .. */
158 /*     .. */
159 /*     .. External Functions .. */
160 /*     .. */
161 /*     .. External Subroutines .. */
162 /*     .. */
163 /*     .. Intrinsic Functions .. */
164 /*     .. */
165 
166 /*     Test the input parameters. */
167 
168     /* Parameter adjustments */
169     a_dim1 = *lda;
170     a_offset = 1 + a_dim1;
171     a -= a_offset;
172     --x;
173     --y;
174 
175     /* Function Body */
176     info = 0;
177     if (! lsame_(uplo, "U", (ftnlen)1, (ftnlen)1) && ! lsame_(uplo, "L", (
178 	    ftnlen)1, (ftnlen)1)) {
179 	info = 1;
180     } else if (*n < 0) {
181 	info = 2;
182     } else if (*k < 0) {
183 	info = 3;
184     } else if (*lda < *k + 1) {
185 	info = 6;
186     } else if (*incx == 0) {
187 	info = 8;
188     } else if (*incy == 0) {
189 	info = 11;
190     }
191     if (info != 0) {
192 	xerbla_("DSBMV ", &info, (ftnlen)6);
193 	return 0;
194     }
195 
196 /*     Quick return if possible. */
197 
198     if (*n == 0 || (*alpha == 0. && *beta == 1.)) {
199 	return 0;
200     }
201 
202 /*     Set up the start points in  X  and  Y. */
203 
204     if (*incx > 0) {
205 	kx = 1;
206     } else {
207 	kx = 1 - (*n - 1) * *incx;
208     }
209     if (*incy > 0) {
210 	ky = 1;
211     } else {
212 	ky = 1 - (*n - 1) * *incy;
213     }
214 
215 /*     Start the operations. In this version the elements of the array A */
216 /*     are accessed sequentially with one pass through A. */
217 
218 /*     First form  y := beta*y. */
219 
220     if (*beta != 1.) {
221 	if (*incy == 1) {
222 	    if (*beta == 0.) {
223 		i__1 = *n;
224 		for (i__ = 1; i__ <= i__1; ++i__) {
225 		    y[i__] = 0.;
226 /* L10: */
227 		}
228 	    } else {
229 		i__1 = *n;
230 		for (i__ = 1; i__ <= i__1; ++i__) {
231 		    y[i__] = *beta * y[i__];
232 /* L20: */
233 		}
234 	    }
235 	} else {
236 	    iy = ky;
237 	    if (*beta == 0.) {
238 		i__1 = *n;
239 		for (i__ = 1; i__ <= i__1; ++i__) {
240 		    y[iy] = 0.;
241 		    iy += *incy;
242 /* L30: */
243 		}
244 	    } else {
245 		i__1 = *n;
246 		for (i__ = 1; i__ <= i__1; ++i__) {
247 		    y[iy] = *beta * y[iy];
248 		    iy += *incy;
249 /* L40: */
250 		}
251 	    }
252 	}
253     }
254     if (*alpha == 0.) {
255 	return 0;
256     }
257     if (lsame_(uplo, "U", (ftnlen)1, (ftnlen)1)) {
258 
259 /*        Form  y  when upper triangle of A is stored. */
260 
261 	kplus1 = *k + 1;
262 	if (*incx == 1 && *incy == 1) {
263 	    i__1 = *n;
264 	    for (j = 1; j <= i__1; ++j) {
265 		temp1 = *alpha * x[j];
266 		temp2 = 0.;
267 		l = kplus1 - j;
268 /* Computing MAX */
269 		i__2 = 1, i__3 = j - *k;
270 		i__4 = j - 1;
271 		for (i__ = max(i__2,i__3); i__ <= i__4; ++i__) {
272 		    y[i__] += temp1 * a[l + i__ + j * a_dim1];
273 		    temp2 += a[l + i__ + j * a_dim1] * x[i__];
274 /* L50: */
275 		}
276 		y[j] = y[j] + temp1 * a[kplus1 + j * a_dim1] + *alpha * temp2;
277 /* L60: */
278 	    }
279 	} else {
280 	    jx = kx;
281 	    jy = ky;
282 	    i__1 = *n;
283 	    for (j = 1; j <= i__1; ++j) {
284 		temp1 = *alpha * x[jx];
285 		temp2 = 0.;
286 		ix = kx;
287 		iy = ky;
288 		l = kplus1 - j;
289 /* Computing MAX */
290 		i__4 = 1, i__2 = j - *k;
291 		i__3 = j - 1;
292 		for (i__ = max(i__4,i__2); i__ <= i__3; ++i__) {
293 		    y[iy] += temp1 * a[l + i__ + j * a_dim1];
294 		    temp2 += a[l + i__ + j * a_dim1] * x[ix];
295 		    ix += *incx;
296 		    iy += *incy;
297 /* L70: */
298 		}
299 		y[jy] = y[jy] + temp1 * a[kplus1 + j * a_dim1] + *alpha *
300 			temp2;
301 		jx += *incx;
302 		jy += *incy;
303 		if (j > *k) {
304 		    kx += *incx;
305 		    ky += *incy;
306 		}
307 /* L80: */
308 	    }
309 	}
310     } else {
311 
312 /*        Form  y  when lower triangle of A is stored. */
313 
314 	if (*incx == 1 && *incy == 1) {
315 	    i__1 = *n;
316 	    for (j = 1; j <= i__1; ++j) {
317 		temp1 = *alpha * x[j];
318 		temp2 = 0.;
319 		y[j] += temp1 * a[j * a_dim1 + 1];
320 		l = 1 - j;
321 /* Computing MIN */
322 		i__4 = *n, i__2 = j + *k;
323 		i__3 = min(i__4,i__2);
324 		for (i__ = j + 1; i__ <= i__3; ++i__) {
325 		    y[i__] += temp1 * a[l + i__ + j * a_dim1];
326 		    temp2 += a[l + i__ + j * a_dim1] * x[i__];
327 /* L90: */
328 		}
329 		y[j] += *alpha * temp2;
330 /* L100: */
331 	    }
332 	} else {
333 	    jx = kx;
334 	    jy = ky;
335 	    i__1 = *n;
336 	    for (j = 1; j <= i__1; ++j) {
337 		temp1 = *alpha * x[jx];
338 		temp2 = 0.;
339 		y[jy] += temp1 * a[j * a_dim1 + 1];
340 		l = 1 - j;
341 		ix = jx;
342 		iy = jy;
343 /* Computing MIN */
344 		i__4 = *n, i__2 = j + *k;
345 		i__3 = min(i__4,i__2);
346 		for (i__ = j + 1; i__ <= i__3; ++i__) {
347 		    ix += *incx;
348 		    iy += *incy;
349 		    y[iy] += temp1 * a[l + i__ + j * a_dim1];
350 		    temp2 += a[l + i__ + j * a_dim1] * x[ix];
351 /* L110: */
352 		}
353 		y[jy] += *alpha * temp2;
354 		jx += *incx;
355 		jy += *incy;
356 /* L120: */
357 	    }
358 	}
359     }
360 
361     return 0;
362 
363 /*     End of DSBMV . */
364 
365 } /* dsbmv_ */
366 
367