File: wsvdc.f

package info (click to toggle)
scilab 4.0-12
  • links: PTS
  • area: non-free
  • in suites: etch, etch-m68k
  • size: 100,640 kB
  • ctags: 57,333
  • sloc: ansic: 377,889; fortran: 242,862; xml: 179,819; tcl: 42,062; sh: 10,593; ml: 9,441; makefile: 4,377; cpp: 1,354; java: 621; csh: 260; yacc: 247; perl: 130; lex: 126; asm: 72; lisp: 30
file content (596 lines) | stat: -rw-r--r-- 19,618 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
      subroutine wsvdc(xr,xi,ldx,n,p,sr,si,er,ei,ur,ui,ldu,vr,vi,ldv,
     *     workr,worki,job,info)
      integer ldx,n,p,ldu,ldv,job,info
      double precision xr(ldx,*),xi(ldx,*),sr(1),si(1),er(1),ei(1),
     *                 ur(ldu,*),ui(ldu,*),vr(ldv,*),vi(ldv,*),
     *                 workr(1),worki(1)
c!purpose
c
c
c     wsvdc is a subroutine to reduce a double-complex nxp matrix x by
c     unitary transformations u and v to diagonal form.  the
c     diagonal elements s(i) are the singular values of x.  the
c     columns of u are the corresponding left singular vectors,
c     and the columns of v the right singular vectors.
c
c!calling sequence
c
c      subroutine wsvdc(xr,xi,ldx,n,p,sr,si,er,ei,ur,ui,ldu,vr,vi,ldv,
c     on entry
c
c         x         double-complex(ldx,p), where ldx.ge.n.
c                   x contains the matrix whose singular value
c                   decomposition is to be computed.  x is
c                   destroyed by wsvdc.
c
c         n         integer.
c                   n is the number of rows of the matrix x.
c
c         p         integer.
c                   p is the number of columns of the matrix x.
c
c         ldx       integer.
c                   ldx is the leading dimension of the array x.
c
c         ldu       integer.
c                   ldu is the leading dimension of the array u
c                   (see below).
c
c         ldv       integer.
c                   ldv is the leading dimension of the array v
c                   (see below).
c
c         work      double-complex(n).
c                   work is a scratch array.
c
c         job       integer.
c                   job controls the computation of the singular
c                   vectors.  it has the decimal expansion ab
c                   with the following meaning
c
c                        a.eq.0    do not compute the left singular
c                                  vectors.
c                        a.eq.1    return the n left singular vectors
c                                  in u.
c                        a.ge.2    returns the first min(n,p)
c                                  left singular vectors in u.
c                        b.eq.0    do not compute the right singular
c                                  vectors.
c                        b.eq.1    return the right singular vectors
c                                  in v.
c
c     on return
c
c         s         double-complex(mm), where mm=min(n+1,p).
c                   the first min(n,p) entries of s contain the
c                   singular values of x arranged in descending
c                   order of magnitude.
c
c         e         double-complex(p).
c                   e ordinarily contains zeros.  however see the
c                   discussion of info for exceptions.
c
c         u         double-complex(ldu,k), where ldu.ge.n.
c                   if joba.eq.1 then k.eq.n,
c                   if joba.eq.2 then k.eq.min(n,p).
c                   u contains the matrix of right singular vectors.
c                   u is not referenced if joba.eq.0.  if n.le.p
c                   or if joba.gt.2, then u may be identified with x
c                   in the subroutine call.
c
c         v         double-complex(ldv,p), where ldv.ge.p.
c                   v contains the matrix of right singular vectors.
c                   v is not referenced if jobb.eq.0.  if p.le.n,
c                   then v may be identified whth x in the
c                   subroutine call.
c
c         info      integer.
c                   the singular values (and their corresponding
c                   singular vectors) s(info+1),s(info+2),...,s(m)
c                   are correct (here m=min(n,p)).  thus if
c                   info.eq.0, all the singular values and their
c                   vectors are correct.  in any event, the matrix
c                   b = ctrans(u)*x*v is the bidiagonal matrix
c                   with the elements of s on its diagonal and the
c                   elements of e on its super-diagonal (ctrans(u)
c                   is the conjugate-transpose of u).  thus the
c                   singular values of x and b are the same.
c
c!originator
c     linpack. this version dated 07/03/79 .
c     g.w. stewart, university of maryland, argonne national lab.
c
c!auxiliary routines
c
c     blas waxpy,pythag,wdotcr,wdotci,wscal,wswap,wnrm2,drotg
c     fortran abs,dimag,max
c     fortran max,min,mod,sqrt
c
c!
c     Copyright INRIA
c     internal variables
c
      integer i,iter,j,jobu,k,kase,kk,l,ll,lls,lm1,lp1,ls,lu,m,maxit,
     *        mm,mm1,mp1,nct,nctp1,ncu,nrt,nrtp1
      double precision pythag,wdotcr,wdotci,tr,ti,rr,ri
      double precision b,c,cs,el,emm1,f,g,wnrm2,scale,shift,sl,sm,sn,
     *                 smm1,t1,test,ztest
      logical wantu,wantv
c
      double precision zdumr,zdumi
      double precision cabs1
      cabs1(zdumr,zdumi) = abs(zdumr) + abs(zdumi)
c
c     set the maximum number of iterations.
c     MODIFIED ACCORDING TO EISPACK HQR2
c
      maxit = 30*min(n,p)
c
c
c     determine what is to be computed.
c
      wantu = .false.
      wantv = .false.
      jobu = mod(job,100)/10
      ncu = n
      if (jobu .gt. 1) ncu = min(n,p)
      if (jobu .ne. 0) wantu = .true.
      if (mod(job,10) .ne. 0) wantv = .true.
c
c     reduce x to bidiagonal form, storing the diagonal elements
c     in s and the super-diagonal elements in e.
c
      info = 0
      nct = min(n-1,p)
      nrt = max(0,min(p-2,n))
      lu = max(nct,nrt)
      if (lu .lt. 1) go to 190
      do 180 l = 1, lu
         lp1 = l + 1
         if (l .gt. nct) go to 30
c
c           compute the transformation for the l-th column and
c           place the l-th diagonal in s(l).
c
            sr(l) = wnrm2(n-l+1,xr(l,l),xi(l,l),1)
            si(l) = 0.0d+0
            if (cabs1(sr(l),si(l)) .eq. 0.0d+0) go to 20
               if (cabs1(xr(l,l),xi(l,l)) .eq. 0.0d+0) go to 10
                  call wsign(sr(l),si(l),xr(l,l),xi(l,l),sr(l),si(l))
   10          continue
               call wdiv(1.0d+0,0.0d+0,sr(l),si(l),tr,ti)
               call wscal(n-l+1,tr,ti,xr(l,l),xi(l,l),1)
               xr(l,l) = 1.0d+0 + xr(l,l)
   20       continue
            sr(l) = -sr(l)
            si(l) = -si(l)
   30    continue
         if (p .lt. lp1) go to 60
         do 50 j = lp1, p
            if (l .gt. nct) go to 40
            if (cabs1(sr(l),si(l)) .eq. 0.0d+0) go to 40
c
c              apply the transformation.
c
               tr = -wdotcr(n-l+1,xr(l,l),xi(l,l),1,xr(l,j),xi(l,j),1)
               ti = -wdotci(n-l+1,xr(l,l),xi(l,l),1,xr(l,j),xi(l,j),1)
               call wdiv(tr,ti,xr(l,l),xi(l,l),tr,ti)
               call waxpy(n-l+1,tr,ti,xr(l,l),xi(l,l),1,xr(l,j),
     *                    xi(l,j),1)
   40       continue
c
c           place the l-th row of x into  e for the
c           subsequent calculation of the row transformation.
c
            er(j) = xr(l,j)
            ei(j) = -xi(l,j)
   50    continue
   60    continue
         if (.not.wantu .or. l .gt. nct) go to 80
c
c           place the transformation in u for subsequent back
c           multiplication.
c
            do 70 i = l, n
               ur(i,l) = xr(i,l)
               ui(i,l) = xi(i,l)
   70       continue
   80    continue
         if (l .gt. nrt) go to 170
c
c           compute the l-th row transformation and place the
c           l-th super-diagonal in e(l).
c
            er(l) = wnrm2(p-l,er(lp1),ei(lp1),1)
            ei(l) = 0.0d+0
            if (cabs1(er(l),ei(l)) .eq. 0.0d+0) go to 100
               if (cabs1(er(lp1),ei(lp1)) .eq. 0.0d+0) go to 90
                  call wsign(er(l),ei(l),er(lp1),ei(lp1),er(l),ei(l))
   90          continue
               call wdiv(1.0d+0,0.0d+0,er(l),ei(l),tr,ti)
               call wscal(p-l,tr,ti,er(lp1),ei(lp1),1)
               er(lp1) = 1.0d+0 + er(lp1)
  100       continue
            er(l) = -er(l)
            ei(l) = +ei(l)
            if (lp1 .gt. n .or. cabs1(er(l),ei(l)) .eq. 0.0d+0)
     *         go to 140
c
c              apply the transformation.
c
               do 110 i = lp1, n
                  workr(i) = 0.0d+0
                  worki(i) = 0.0d+0
  110          continue
               do 120 j = lp1, p
                  call waxpy(n-l,er(j),ei(j),xr(lp1,j),xi(lp1,j),1,
     *                       workr(lp1),worki(lp1),1)
  120          continue
               do 130 j = lp1, p
                  call wdiv(-er(j),-ei(j),er(lp1),ei(lp1),tr,ti)
                  call waxpy(n-l,tr,-ti,workr(lp1),worki(lp1),1,
     *                       xr(lp1,j),xi(lp1,j),1)
  130          continue
  140       continue
            if (.not.wantv) go to 160
c
c              place the transformation in v for subsequent
c              back multiplication.
c
               do 150 i = lp1, p
                  vr(i,l) = er(i)
                  vi(i,l) = ei(i)
  150          continue
  160       continue
  170    continue
  180 continue
  190 continue
c
c     set up the final bidiagonal matrix or order m.
c
      m = min(p,n+1)
      nctp1 = nct + 1
      nrtp1 = nrt + 1
      if (nct .ge. p) go to 200
         sr(nctp1) = xr(nctp1,nctp1)
         si(nctp1) = xi(nctp1,nctp1)
  200 continue
      if (n .ge. m) go to 210
         sr(m) = 0.0d+0
         si(m) = 0.0d+0
  210 continue
      if (nrtp1 .ge. m) go to 220
         er(nrtp1) = xr(nrtp1,m)
         ei(nrtp1) = xi(nrtp1,m)
  220 continue
      er(m) = 0.0d+0
      ei(m) = 0.0d+0
c
c     if required, generate u.
c
      if (.not.wantu) go to 350
         if (ncu .lt. nctp1) go to 250
         do 240 j = nctp1, ncu
            do 230 i = 1, n
               ur(i,j) = 0.0d+0
               ui(i,j) = 0.0d+0
  230       continue
            ur(j,j) = 1.0d+0
            ui(j,j) = 0.0d+0
  240    continue
  250    continue
         if (nct .lt. 1) go to 340
         do 330 ll = 1, nct
            l = nct - ll + 1
            if (cabs1(sr(l),si(l)) .eq. 0.0d+0) go to 300
               lp1 = l + 1
               if (ncu .lt. lp1) go to 270
               do 260 j = lp1, ncu
                  tr = -wdotcr(n-l+1,ur(l,l),ui(l,l),1,ur(l,j),
     *                         ui(l,j),1)
                  ti = -wdotci(n-l+1,ur(l,l),ui(l,l),1,ur(l,j),
     *                         ui(l,j),1)
                  call wdiv(tr,ti,ur(l,l),ui(l,l),tr,ti)
                  call waxpy(n-l+1,tr,ti,ur(l,l),ui(l,l),1,ur(l,j),
     *                       ui(l,j),1)
  260          continue
  270          continue
               call wrscal(n-l+1,-1.0d+0,ur(l,l),ui(l,l),1)
               ur(l,l) = 1.0d+0 + ur(l,l)
               lm1 = l - 1
               if (lm1 .lt. 1) go to 290
               do 280 i = 1, lm1
                  ur(i,l) = 0.0d+0
                  ui(i,l) = 0.0d+0
  280          continue
  290          continue
            go to 320
  300       continue
               do 310 i = 1, n
                  ur(i,l) = 0.0d+0
                  ui(i,l) = 0.0d+0
  310          continue
               ur(l,l) = 1.0d+0
               ui(l,l) = 0.0d+0
  320       continue
  330    continue
  340    continue
  350 continue
c
c     if it is required, generate v.
c
      if (.not.wantv) go to 400
         do 390 ll = 1, p
            l = p - ll + 1
            lp1 = l + 1
            if (l .gt. nrt) go to 370
            if (cabs1(er(l),ei(l)) .eq. 0.0d+0) go to 370
               do 360 j = lp1, p
                  tr = -wdotcr(p-l,vr(lp1,l),vi(lp1,l),1,vr(lp1,j),
     *                         vi(lp1,j),1)
                  ti = -wdotci(p-l,vr(lp1,l),vi(lp1,l),1,vr(lp1,j),
     *                         vi(lp1,j),1)
                  call wdiv(tr,ti,vr(lp1,l),vi(lp1,l),tr,ti)
                  call waxpy(p-l,tr,ti,vr(lp1,l),vi(lp1,l),1,vr(lp1,j),
     *                       vi(lp1,j),1)
  360          continue
  370       continue
            do 380 i = 1, p
               vr(i,l) = 0.0d+0
               vi(i,l) = 0.0d+0
  380       continue
            vr(l,l) = 1.0d+0
            vi(l,l) = 0.0d+0
  390    continue
  400 continue
c
c     transform s and e so that they are real.
c
      do 420 i = 1, m
            tr = pythag(sr(i),si(i))
            if (tr .eq. 0.0d+0) go to 405
            rr = sr(i)/tr
            ri = si(i)/tr
            sr(i) = tr
            si(i) = 0.0d+0
            if (i .lt. m) call wdiv(er(i),ei(i),rr,ri,er(i),ei(i))
            if (wantu) call wscal(n,rr,ri,ur(1,i),ui(1,i),1)
  405    continue
c     ...exit
         if (i .eq. m) go to 430
            tr = pythag(er(i),ei(i))
            if (tr .eq. 0.0d+0) go to 410
            call wdiv(tr,0.0d+0,er(i),ei(i),rr,ri)
            er(i) = tr
            ei(i) = 0.0d+0
            call wmul(sr(i+1),si(i+1),rr,ri,sr(i+1),si(i+1))
            if (wantv) call wscal(p,rr,ri,vr(1,i+1),vi(1,i+1),1)
  410    continue
  420 continue
  430 continue
c
c     main iteration loop for the singular values.
c
      mm = m
      iter = 0
  440 continue
c
c        quit if all the singular values have been found.
c
c     ...exit
         if (m .eq. 0) go to 700
c
c        if too many iterations have been performed, set
c        flag and return.
c
         if (iter .lt. maxit) go to 450
            info = m
c     ......exit
            go to 700
  450    continue
c
c        this section of the program inspects for
c        negligible elements in the s and e arrays.  on
c        completion the variable kase is set as follows.
c
c           kase = 1     if sr(m) and er(l-1) are negligible and l.lt.m
c           kase = 2     if sr(l) is negligible and l.lt.m
c           kase = 3     if er(l-1) is negligible, l.lt.m, and
c                        sr(l), ..., sr(m) are not negligible (qr step).
c           kase = 4     if er(m-1) is negligible (convergence).
c
         do 470 ll = 1, m
            l = m - ll
c        ...exit
            if (l .eq. 0) go to 480
            test = pythag(sr(l),si(l)) + pythag(sr(l+1),si(l+1))
            ztest = test + pythag(er(l),ei(l))
            if (ztest .ne. test) go to 460
               er(l) = 0.0d+0
               ei(l) = 0.0d+0
c        ......exit
               go to 480
  460       continue
  470    continue
  480    continue
         if (l .ne. m - 1) go to 490
            kase = 4
         go to 560
  490    continue
            lp1 = l + 1
            mp1 = m + 1
            do 510 lls = lp1, mp1
               ls = m - lls + lp1
c           ...exit
               if (ls .eq. l) go to 520
               test = 0.0d+0
               if (ls .ne. m) test=test + pythag(er(ls),ei(ls))
               if (ls .ne. l + 1) test=test + pythag(er(ls-1),ei(ls-1))
               ztest = test + pythag(sr(ls),si(ls))
               if (ztest .ne. test) go to 500
                  sr(ls) = 0.0d+0
                  si(ls) = 0.0d+0
c           ......exit
                  go to 520
  500          continue
  510       continue
  520       continue
            if (ls .ne. l) go to 530
               kase = 3
            go to 550
  530       continue
            if (ls .ne. m) go to 540
               kase = 1
            go to 550
  540       continue
               kase = 2
               l = ls
  550       continue
  560    continue
         l = l + 1
c
c        perform the task indicated by kase.
c
         go to (570, 600, 620, 650), kase
c
c        deflate negligible s(m).
c
  570    continue
            mm1 = m - 1
            f = er(m-1)
            er(m-1) = 0.0d+0
            ei(m-1) = 0.0d+0
            do 590 kk = l, mm1
               k = mm1 - kk + l
               t1 = sr(k)
               call drotg(t1,f,cs,sn)
               sr(k) = t1
               si(k) = 0.0d0
               if (k .eq. l) go to 580
                  f = -sn*er(k-1)
                  er(k-1) = cs*er(k-1)
                  ei(k-1) = cs*ei(k-1)
  580          continue
               if (wantv) call drot(p,vr(1,k),1,vr(1,m),1,cs,sn)
               if (wantv) call drot(p,vi(1,k),1,vi(1,m),1,cs,sn)
  590       continue
         go to 690
c
c        split at negligible s(l).
c
  600    continue
            f = er(l-1)
            er(l-1) = 0.0d+0
            ei(l-1) = 0.0d+0
            do 610 k = l, m
               t1 = sr(k)
               call drotg(t1,f,cs,sn)
               sr(k) = t1
               si(k) = 0.0d0
               f = -sn*er(k)
               er(k) = cs*er(k)
               ei(k) = cs*ei(k)
               if (wantu) call drot(n,ur(1,k),1,ur(1,l-1),1,cs,sn)
               if (wantu) call drot(n,ui(1,k),1,ui(1,l-1),1,cs,sn)
  610       continue
         go to 690
c
c        perform one qr step.
c
  620    continue
c
c           calculate the shift.
c
            scale = max(pythag(sr(m),si(m)),pythag(sr(m-1),si(m-1)),
     *        pythag(er(m-1),ei(m-1)),
     *        pythag(sr(l),si(l)),pythag(er(l),ei(l)))
            sm = sr(m)/scale
            smm1 = sr(m-1)/scale
            emm1 = er(m-1)/scale
            sl = sr(l)/scale
            el = er(l)/scale
            b = ((smm1 + sm)*(smm1 - sm) + emm1**2)/2.0d+0
            c = (sm*emm1)**2
            shift = 0.0d+0
            if (b .eq. 0.0d+0 .and. c .eq. 0.0d+0) go to 630
               shift = sqrt(b**2+c)
               if (b .lt. 0.0d+0) shift = -shift
               shift = c/(b + shift)
  630       continue
c            f = (sl + sm)*(sl - sm) - shift
        f=(sl+sm)*(sl-sm)+shift
            g = sl*el
c
c           chase zeros.
c
            mm1 = m - 1
            do 640 k = l, mm1
               call drotg(f,g,cs,sn)
               if (k .ne. l) then
                  er(k-1) = f
                  ei(k-1) = 0.0d0
               endif
               f = cs*sr(k) + sn*er(k)
               er(k) = cs*er(k) - sn*sr(k)
               ei(k) = cs*ei(k) - sn*si(k)
               g = sn*sr(k+1)
               sr(k+1) = cs*sr(k+1)
               si(k+1) = cs*si(k+1)
               if (wantv) call drot(p,vr(1,k),1,vr(1,k+1),1,cs,sn)
               if (wantv) call drot(p,vi(1,k),1,vi(1,k+1),1,cs,sn)
               call drotg(f,g,cs,sn)
               sr(k) = f
               si(k) = 0.0d0
               f = cs*er(k) + sn*sr(k+1)
               sr(k+1) = -sn*er(k) + cs*sr(k+1)
               si(k+1) = -sn*ei(k) + cs*si(k+1)
               g = sn*er(k+1)
               er(k+1) = cs*er(k+1)
               ei(k+1) = cs*ei(k+1)
               if (wantu .and. k .lt. n)
     *            call drot(n,ur(1,k),1,ur(1,k+1),1,cs,sn)
               if (wantu .and. k .lt. n)
     *            call drot(n,ui(1,k),1,ui(1,k+1),1,cs,sn)
  640       continue
            er(m-1) = f
            ei(m-1) = 0.0d0
            iter = iter + 1
         go to 690
c
c        convergence
c
  650    continue
c
c           make the singular value  positive
c
            if (sr(l) .ge. 0.0d+0) go to 660
               sr(l) = -sr(l)
               si(l) = -si(l)
             if (wantv) call wrscal(p,-1.0d+0,vr(1,l),vi(1,l),1)
  660       continue
c
c           order the singular value.
c
  670       if (l .eq. mm) go to 680
c           ...exit
               if (sr(l) .ge. sr(l+1)) go to 680
               tr = sr(l)
               sr(l) = sr(l+1)
               sr(l+1) = tr
               tr = si(l)
               si(l) = si(l+1)
               si(l+1) = tr
               if (wantv .and. l .lt. p)
     *            call wswap(p,vr(1,l),vi(1,l),1,vr(1,l+1),vi(1,l+1),1)
               if (wantu .and. l .lt. n)
     *            call wswap(n,ur(1,l),ui(1,l),1,ur(1,l+1),ui(1,l+1),1)
               l = l + 1
            go to 670
  680       continue
            iter = 0
            m = m - 1
  690    continue
      go to 440
  700 continue
      return
      end