MODFLOW 6  version 6.6.0.dev0
USGS Modular Hydrologic Model
blas1_d.f90
Go to the documentation of this file.
1 function dasum ( n, x, incx )
2 
3  !*****************************************************************************80
4  !
5  !! DASUM takes the sum of the absolute values of a vector.
6  !
7  ! Discussion:
8  !
9  ! This routine uses double precision real arithmetic.
10  !
11  ! Licensing:
12  !
13  ! This code is distributed under the GNU LGPL license.
14  !
15  ! Modified:
16  !
17  ! 15 February 2001
18  !
19  ! Author:
20  !
21  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
22  ! David Kincaid, Fred Krogh.
23  ! FORTRAN90 version by John Burkardt.
24  !
25  ! Reference:
26  !
27  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
28  ! LINPACK User's Guide,
29  ! SIAM, 1979,
30  ! ISBN13: 978-0-898711-72-1,
31  ! LC: QA214.L56.
32  !
33  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
34  ! Algorithm 539,
35  ! Basic Linear Algebra Subprograms for Fortran Usage,
36  ! ACM Transactions on Mathematical Software,
37  ! Volume 5, Number 3, September 1979, pages 308-323.
38  !
39  ! Parameters:
40  !
41  ! Input, integer ( kind = 4 ) N, the number of entries in the vector.
42  !
43  ! Input, real ( kind = 8 ) X(*), the vector to be examined.
44  !
45  ! Input, integer ( kind = 4 ) INCX, the increment between successive
46  ! entries of X. INCX must not be negative.
47  !
48  ! Output, real ( kind = 8 ) DASUM, the sum of the absolute values of X.
49  !
50  implicit none
51 
52  real ( kind = 8 ) dasum
53  integer ( kind = 4 ) incx
54  integer ( kind = 4 ) n
55  real ( kind = 8 ) x(*)
56 
57  dasum = sum( abs( x(1:1+(n-1)*incx:incx) ) )
58 
59  return
60  end
61  subroutine daxpy ( n, da, dx, incx, dy, incy )
62 
63  !*****************************************************************************80
64  !
65  !! DAXPY computes constant times a vector plus a vector.
66  !
67  ! Discussion:
68  !
69  ! This routine uses double precision real arithmetic.
70  !
71  ! This routine uses unrolled loops for increments equal to one.
72  !
73  ! Licensing:
74  !
75  ! This code is distributed under the GNU LGPL license.
76  !
77  ! Modified:
78  !
79  ! 16 May 2005
80  !
81  ! Author:
82  !
83  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
84  ! David Kincaid, Fred Krogh.
85  ! FORTRAN90 version by John Burkardt.
86  !
87  ! Reference:
88  !
89  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
90  ! LINPACK User's Guide,
91  ! SIAM, 1979,
92  ! ISBN13: 978-0-898711-72-1,
93  ! LC: QA214.L56.
94  !
95  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
96  ! Algorithm 539,
97  ! Basic Linear Algebra Subprograms for Fortran Usage,
98  ! ACM Transactions on Mathematical Software,
99  ! Volume 5, Number 3, September 1979, pages 308-323.
100  !
101  ! Parameters:
102  !
103  ! Input, integer ( kind = 4 ) N, the number of elements in DX and DY.
104  !
105  ! Input, real ( kind = 8 ) DA, the multiplier of DX.
106  !
107  ! Input, real ( kind = 8 ) DX(*), the first vector.
108  !
109  ! Input, integer ( kind = 4 ) INCX, the increment between successive
110  ! entries of DX.
111  !
112  ! Input/output, real ( kind = 8 ) DY(*), the second vector.
113  ! On output, DY(*) has been replaced by DY(*) + DA * DX(*).
114  !
115  ! Input, integer ( kind = 4 ) INCY, the increment between successive
116  ! entries of DY.
117  !
118  implicit none
119 
120  real ( kind = 8 ) da
121  real ( kind = 8 ) dx(*)
122  real ( kind = 8 ) dy(*)
123  integer ( kind = 4 ) i
124  integer ( kind = 4 ) incx
125  integer ( kind = 4 ) incy
126  integer ( kind = 4 ) ix
127  integer ( kind = 4 ) iy
128  integer ( kind = 4 ) m
129  integer ( kind = 4 ) n
130 
131  if ( n <= 0 ) then
132  return
133  end if
134 
135  if ( da == 0.0d+00 ) then
136  return
137  end if
138  !
139  ! Code for unequal increments or equal increments
140  ! not equal to 1.
141  !
142  if ( incx /= 1 .or. incy /= 1 ) then
143 
144  if ( 0 <= incx ) then
145  ix = 1
146  else
147  ix = ( - n + 1 ) * incx + 1
148  end if
149 
150  if ( 0 <= incy ) then
151  iy = 1
152  else
153  iy = ( - n + 1 ) * incy + 1
154  end if
155 
156  do i = 1, n
157  dy(iy) = dy(iy) + da * dx(ix)
158  ix = ix + incx
159  iy = iy + incy
160  end do
161  !
162  ! Code for both increments equal to 1.
163  !
164  else
165 
166  m = mod( n, 4 )
167 
168  dy(1:m) = dy(1:m) + da * dx(1:m)
169 
170  do i = m + 1, n, 4
171  dy(i ) = dy(i ) + da * dx(i )
172  dy(i+1) = dy(i+1) + da * dx(i+1)
173  dy(i+2) = dy(i+2) + da * dx(i+2)
174  dy(i+3) = dy(i+3) + da * dx(i+3)
175  end do
176 
177  end if
178 
179  return
180  end
181  subroutine dcopy ( n, dx, incx, dy, incy )
182 
183  !*****************************************************************************80
184  !
185  !! DCOPY copies a vector X to a vector Y.
186  !
187  ! Discussion:
188  !
189  ! This routine uses double precision real arithmetic.
190  !
191  ! The routine uses unrolled loops for increments equal to one.
192  !
193  ! Licensing:
194  !
195  ! This code is distributed under the GNU LGPL license.
196  !
197  ! Modified:
198  !
199  ! 16 May 2005
200  !
201  ! Author:
202  !
203  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
204  ! David Kincaid, Fred Krogh.
205  ! FORTRAN90 version by John Burkardt.
206  !
207  ! Reference:
208  !
209  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
210  ! LINPACK User's Guide,
211  ! SIAM, 1979,
212  ! ISBN13: 978-0-898711-72-1,
213  ! LC: QA214.L56.
214  !
215  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
216  ! Algorithm 539,
217  ! Basic Linear Algebra Subprograms for Fortran Usage,
218  ! ACM Transactions on Mathematical Software,
219  ! Volume 5, Number 3, September 1979, pages 308-323.
220  !
221  ! Parameters:
222  !
223  ! Input, integer ( kind = 4 ) N, the number of elements in DX and DY.
224  !
225  ! Input, real ( kind = 8 ) DX(*), the first vector.
226  !
227  ! Input, integer ( kind = 4 ) INCX, the increment between successive
228  ! entries of DX.
229  !
230  ! Output, real ( kind = 8 ) DY(*), the second vector.
231  !
232  ! Input, integer ( kind = 4 ) INCY, the increment between successive
233  ! entries of DY.
234  !
235  implicit none
236 
237  integer ( kind = 4 ) i
238  integer ( kind = 4 ) incx
239  integer ( kind = 4 ) incy
240  integer ( kind = 4 ) ix
241  integer ( kind = 4 ) iy
242  integer ( kind = 4 ) m
243  integer ( kind = 4 ) n
244  real ( kind = 8 ) dx(*)
245  real ( kind = 8 ) dy(*)
246 
247  if ( n <= 0 ) then
248  return
249  end if
250 
251  if ( incx == 1 .and. incy == 1 ) then
252 
253  m = mod( n, 7 )
254 
255  if ( m /= 0 ) then
256 
257  dy(1:m) = dx(1:m)
258 
259  end if
260 
261  do i = m+1, n, 7
262  dy(i) = dx(i)
263  dy(i + 1) = dx(i + 1)
264  dy(i + 2) = dx(i + 2)
265  dy(i + 3) = dx(i + 3)
266  dy(i + 4) = dx(i + 4)
267  dy(i + 5) = dx(i + 5)
268  dy(i + 6) = dx(i + 6)
269  end do
270 
271  else
272 
273  if ( 0 <= incx ) then
274  ix = 1
275  else
276  ix = ( -n + 1 ) * incx + 1
277  end if
278 
279  if ( 0 <= incy ) then
280  iy = 1
281  else
282  iy = ( -n + 1 ) * incy + 1
283  end if
284 
285  do i = 1, n
286  dy(iy) = dx(ix)
287  ix = ix + incx
288  iy = iy + incy
289  end do
290 
291  end if
292 
293  return
294  end
295  function ddot ( n, dx, incx, dy, incy )
296 
297  !*****************************************************************************80
298  !
299  !! DDOT forms the dot product of two vectors.
300  !
301  ! Licensing:
302  !
303  ! This code is distributed under the GNU LGPL license.
304  !
305  ! Modified:
306  !
307  ! 06 June 2014
308  !
309  ! Author:
310  !
311  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
312  ! David Kincaid, Fred Krogh.
313  ! FORTRAN90 version by John Burkardt.
314  !
315  ! Reference:
316  !
317  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
318  ! LINPACK User's Guide,
319  ! SIAM, 1979,
320  ! ISBN13: 978-0-898711-72-1,
321  ! LC: QA214.L56.
322  !
323  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
324  ! Algorithm 539,
325  ! Basic Linear Algebra Subprograms for Fortran Usage,
326  ! ACM Transactions on Mathematical Software,
327  ! Volume 5, Number 3, September 1979, pages 308-323.
328  !
329  ! Parameters:
330  !
331  ! Input, integer ( kind = 4 ) N, the number of entries in the vectors.
332  !
333  ! Input, real ( kind = 8 ) DX(*), the first vector.
334  !
335  ! Input, integer ( kind = 4 ) INCX, the increment between successive
336  ! entries in DX.
337  !
338  ! Input, real ( kind = 8 ) DY(*), the second vector.
339  !
340  ! Input, integer ( kind = 4 ) INCY, the increment between successive
341  ! entries in DY.
342  !
343  ! Output, real ( kind = 8 ) DDOT, the sum of the product of the
344  ! corresponding entries of DX and DY.
345  !
346  implicit none
347 
348  real ( kind = 8 ) ddot
349  real ( kind = 8 ) dx(*)
350  real ( kind = 8 ) dy(*)
351  integer ( kind = 4 ) incx
352  integer ( kind = 4 ) incy
353  integer ( kind = 4 ) n
354  integer ( kind = 4 ) x1
355  integer ( kind = 4 ) xi
356  integer ( kind = 4 ) xn
357  integer ( kind = 4 ) y1
358  integer ( kind = 4 ) yi
359  integer ( kind = 4 ) yn
360  !
361  ! Let the FORTRAN90 function DOT_PRODUCT take care of optimization.
362  !
363  if ( 0 < incx ) then
364  x1 = 1
365  xn = 1 + ( n - 1 ) * incx
366  xi = incx
367  else
368  x1 = 1 + ( n - 1 ) * incx
369  xn = 1
370  xi = - incx
371  end if
372 
373  if ( 0 < incy ) then
374  y1 = 1
375  yn = 1 + ( n - 1 ) * incy
376  yi = incy
377  else
378  y1 = 1 + ( n - 1 ) * incy
379  yn = 1
380  yi = - incy
381  end if
382 
383  ddot = dot_product( dx(x1:xn:xi), dy(y1:yn:yi) )
384 
385  return
386  end
387  function dnrm2 ( n, x, incx )
388 
389  !*****************************************************************************80
390  !
391  !! DNRM2 returns the euclidean norm of a vector.
392  !
393  ! Discussion:
394  !
395  ! This routine uses double precision real arithmetic.
396  !
397  ! DNRM2 ( X ) = sqrt ( X' * X )
398  !
399  ! Licensing:
400  !
401  ! This code is distributed under the GNU LGPL license.
402  !
403  ! Modified:
404  !
405  ! 16 May 2005
406  !
407  ! Author:
408  !
409  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
410  ! David Kincaid, Fred Krogh.
411  ! FORTRAN90 version by John Burkardt.
412  !
413  ! Reference:
414  !
415  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
416  ! LINPACK User's Guide,
417  ! SIAM, 1979,
418  ! ISBN13: 978-0-898711-72-1,
419  ! LC: QA214.L56.
420  !
421  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
422  ! Algorithm 539,
423  ! Basic Linear Algebra Subprograms for Fortran Usage,
424  ! ACM Transactions on Mathematical Software,
425  ! Volume 5, Number 3, September 1979, pages 308-323.
426  !
427  ! Parameters:
428  !
429  ! Input, integer ( kind = 4 ) N, the number of entries in the vector.
430  !
431  ! Input, real ( kind = 8 ) X(*), the vector whose norm is to be computed.
432  !
433  ! Input, integer ( kind = 4 ) INCX, the increment between successive
434  ! entries of X.
435  !
436  ! Output, real ( kind = 8 ) DNRM2, the Euclidean norm of X.
437  !
438  implicit none
439 
440  real ( kind = 8 ) absxi
441  real ( kind = 8 ) dnrm2
442  integer ( kind = 4 ) incx
443  integer ( kind = 4 ) ix
444  integer ( kind = 4 ) n
445  real ( kind = 8 ) norm
446  real ( kind = 8 ) scale
447  real ( kind = 8 ) ssq
448  real ( kind = 8 ) x(*)
449 
450  if ( n < 1 .or. incx < 1 ) then
451 
452  norm = 0.0d+00
453 
454  else if ( n == 1 ) then
455 
456  norm = abs( x(1) )
457 
458  else
459 
460  scale = 0.0d+00
461  ssq = 1.0d+00
462 
463  do ix = 1, 1 + ( n - 1 ) * incx, incx
464  if ( x(ix) /= 0.0d+00 ) then
465  absxi = abs( x(ix) )
466  if ( scale < absxi ) then
467  ssq = 1.0d+00 + ssq * ( scale / absxi )**2
468  scale = absxi
469  else
470  ssq = ssq + ( absxi / scale )**2
471  end if
472  end if
473  end do
474  norm = scale * sqrt( ssq )
475  end if
476 
477  dnrm2 = norm
478 
479  return
480  end
481  subroutine drot ( n, x, incx, y, incy, c, s )
482 
483  !*****************************************************************************80
484  !
485  !! DROT applies a plane rotation.
486  !
487  ! Discussion:
488  !
489  ! This routine uses double precision real arithmetic.
490  !
491  ! Licensing:
492  !
493  ! This code is distributed under the GNU LGPL license.
494  !
495  ! Modified:
496  !
497  ! 08 April 1999
498  !
499  ! Author:
500  !
501  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
502  ! David Kincaid, Fred Krogh.
503  ! FORTRAN90 version by John Burkardt.
504  !
505  ! Reference:
506  !
507  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
508  ! LINPACK User's Guide,
509  ! SIAM, 1979,
510  ! ISBN13: 978-0-898711-72-1,
511  ! LC: QA214.L56.
512  !
513  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
514  ! Algorithm 539,
515  ! Basic Linear Algebra Subprograms for Fortran Usage,
516  ! ACM Transactions on Mathematical Software,
517  ! Volume 5, Number 3, September 1979, pages 308-323.
518  !
519  ! Parameters:
520  !
521  ! Input, integer ( kind = 4 ) N, the number of entries in the vectors.
522  !
523  ! Input/output, real ( kind = 8 ) X(*), one of the vectors to be rotated.
524  !
525  ! Input, integer ( kind = 4 ) INCX, the increment between successive
526  ! entries of X.
527  !
528  ! Input/output, real ( kind = 8 ) Y(*), one of the vectors to be rotated.
529  !
530  ! Input, integer ( kind = 4 ) INCY, the increment between successive
531  ! elements of Y.
532  !
533  ! Input, real ( kind = 8 ) C, S, parameters (presumably the cosine and
534  ! sine of some angle) that define a plane rotation.
535  !
536  implicit none
537 
538  real ( kind = 8 ) c
539  integer ( kind = 4 ) i
540  integer ( kind = 4 ) incx
541  integer ( kind = 4 ) incy
542  integer ( kind = 4 ) ix
543  integer ( kind = 4 ) iy
544  integer ( kind = 4 ) n
545  real ( kind = 8 ) s
546  real ( kind = 8 ) stemp
547  real ( kind = 8 ) x(*)
548  real ( kind = 8 ) y(*)
549 
550  if ( n <= 0 ) then
551 
552  else if ( incx == 1 .and. incy == 1 ) then
553 
554  do i = 1, n
555  stemp = c * x(i) + s * y(i)
556  y(i) = c * y(i) - s * x(i)
557  x(i) = stemp
558  end do
559 
560  else
561 
562  if ( 0 <= incx ) then
563  ix = 1
564  else
565  ix = ( - n + 1 ) * incx + 1
566  end if
567 
568  if ( 0 <= incy ) then
569  iy = 1
570  else
571  iy = ( - n + 1 ) * incy + 1
572  end if
573 
574  do i = 1, n
575  stemp = c * x(ix) + s * y(iy)
576  y(iy) = c * y(iy) - s * x(ix)
577  x(ix) = stemp
578  ix = ix + incx
579  iy = iy + incy
580  end do
581 
582  end if
583 
584  return
585  end
586  subroutine drotg ( sa, sb, c, s )
587 
588  !*****************************************************************************80
589  !
590  !! DROTG constructs a Givens plane rotation.
591  !
592  ! Discussion:
593  !
594  ! Given values A and B, this routine computes
595  !
596  ! SIGMA = sign ( A ) if abs ( A ) > abs ( B )
597  ! = sign ( B ) if abs ( A ) <= abs ( B );
598  !
599  ! R = SIGMA * ( A * A + B * B );
600  !
601  ! C = A / R if R is not 0
602  ! = 1 if R is 0;
603  !
604  ! S = B / R if R is not 0,
605  ! 0 if R is 0.
606  !
607  ! The computed numbers then satisfy the equation
608  !
609  ! ( C S ) ( A ) = ( R )
610  ! ( -S C ) ( B ) = ( 0 )
611  !
612  ! The routine also computes
613  !
614  ! Z = S if abs ( A ) > abs ( B ),
615  ! = 1 / C if abs ( A ) <= abs ( B ) and C is not 0,
616  ! = 1 if C is 0.
617  !
618  ! The single value Z encodes C and S, and hence the rotation:
619  !
620  ! If Z = 1, set C = 0 and S = 1;
621  ! If abs ( Z ) < 1, set C = sqrt ( 1 - Z * Z ) and S = Z;
622  ! if abs ( Z ) > 1, set C = 1/ Z and S = sqrt ( 1 - C * C );
623  !
624  ! Licensing:
625  !
626  ! This code is distributed under the GNU LGPL license.
627  !
628  ! Modified:
629  !
630  ! 15 May 2006
631  !
632  ! Author:
633  !
634  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
635  ! David Kincaid, Fred Krogh.
636  ! FORTRAN90 version by John Burkardt.
637  !
638  ! Reference:
639  !
640  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
641  ! LINPACK User's Guide,
642  ! SIAM, 1979,
643  ! ISBN13: 978-0-898711-72-1,
644  ! LC: QA214.L56.
645  !
646  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
647  ! Algorithm 539,
648  ! Basic Linear Algebra Subprograms for Fortran Usage,
649  ! ACM Transactions on Mathematical Software,
650  ! Volume 5, Number 3, September 1979, pages 308-323.
651  !
652  ! Parameters:
653  !
654  ! Input/output, real ( kind = 8 ) SA, SB. On input, SA and SB are the values
655  ! A and B. On output, SA is overwritten with R, and SB is
656  ! overwritten with Z.
657  !
658  ! Output, real ( kind = 8 ) C, S, the cosine and sine of the
659  ! Givens rotation.
660  !
661  implicit none
662 
663  real ( kind = 8 ) c
664  real ( kind = 8 ) r
665  real ( kind = 8 ) roe
666  real ( kind = 8 ) s
667  real ( kind = 8 ) sa
668  real ( kind = 8 ) sb
669  real ( kind = 8 ) scale
670  real ( kind = 8 ) z
671 
672  if ( abs( sb ) < abs( sa ) ) then
673  roe = sa
674  else
675  roe = sb
676  end if
677 
678  scale = abs( sa ) + abs( sb )
679 
680  if ( scale == 0.0d+00 ) then
681  c = 1.0d+00
682  s = 0.0d+00
683  r = 0.0d+00
684  else
685  r = scale * sqrt( ( sa / scale )**2 + ( sb / scale )**2 )
686  r = sign( 1.0d+00, roe ) * r
687  c = sa / r
688  s = sb / r
689  end if
690 
691  if ( 0.0d+00 < abs( c ) .and. abs( c ) <= s ) then
692  z = 1.0d+00 / c
693  else
694  z = s
695  end if
696 
697  sa = r
698  sb = z
699 
700  return
701  end
702  subroutine dscal ( n, sa, x, incx )
703 
704  !*****************************************************************************80
705  !
706  !! DSCAL scales a vector by a constant.
707  !
708  ! Discussion:
709  !
710  ! This routine uses double precision real arithmetic.
711  !
712  ! Licensing:
713  !
714  ! This code is distributed under the GNU LGPL license.
715  !
716  ! Modified:
717  !
718  ! 08 April 1999
719  !
720  ! Author:
721  !
722  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
723  ! David Kincaid, Fred Krogh.
724  ! FORTRAN90 version by John Burkardt.
725  !
726  ! Reference:
727  !
728  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
729  ! LINPACK User's Guide,
730  ! SIAM, 1979,
731  ! ISBN13: 978-0-898711-72-1,
732  ! LC: QA214.L56.
733  !
734  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
735  ! Algorithm 539,
736  ! Basic Linear Algebra Subprograms for Fortran Usage,
737  ! ACM Transactions on Mathematical Software,
738  ! Volume 5, Number 3, September 1979, pages 308-323.
739  !
740  ! Parameters:
741  !
742  ! Input, integer ( kind = 4 ) N, the number of entries in the vector.
743  !
744  ! Input, real ( kind = 8 ) SA, the multiplier.
745  !
746  ! Input/output, real ( kind = 8 ) X(*), the vector to be scaled.
747  !
748  ! Input, integer ( kind = 4 ) INCX, the increment between successive
749  ! entries of X.
750  !
751  implicit none
752 
753  integer ( kind = 4 ) i
754  integer ( kind = 4 ) incx
755  integer ( kind = 4 ) ix
756  integer ( kind = 4 ) m
757  integer ( kind = 4 ) n
758  real ( kind = 8 ) sa
759  real ( kind = 8 ) x(*)
760 
761  if ( n <= 0 ) then
762 
763  else if ( incx == 1 ) then
764 
765  m = mod( n, 5 )
766 
767  x(1:m) = sa * x(1:m)
768 
769  do i = m+1, n, 5
770  x(i) = sa * x(i)
771  x(i+1) = sa * x(i+1)
772  x(i+2) = sa * x(i+2)
773  x(i+3) = sa * x(i+3)
774  x(i+4) = sa * x(i+4)
775  end do
776 
777  else
778 
779  if ( 0 <= incx ) then
780  ix = 1
781  else
782  ix = ( - n + 1 ) * incx + 1
783  end if
784 
785  do i = 1, n
786  x(ix) = sa * x(ix)
787  ix = ix + incx
788  end do
789 
790  end if
791 
792  return
793  end
794  subroutine dswap ( n, x, incx, y, incy )
795 
796  !*****************************************************************************80
797  !
798  !! DSWAP interchanges two vectors.
799  !
800  ! Discussion:
801  !
802  ! This routine uses double precision real arithmetic.
803  !
804  ! Licensing:
805  !
806  ! This code is distributed under the GNU LGPL license.
807  !
808  ! Modified:
809  !
810  ! 08 April 1999
811  !
812  ! Author:
813  !
814  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
815  ! David Kincaid, Fred Krogh.
816  ! FORTRAN90 version by John Burkardt.
817  !
818  ! Reference:
819  !
820  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
821  ! LINPACK User's Guide,
822  ! SIAM, 1979,
823  ! ISBN13: 978-0-898711-72-1,
824  ! LC: QA214.L56.
825  !
826  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
827  ! Algorithm 539,
828  ! Basic Linear Algebra Subprograms for Fortran Usage,
829  ! ACM Transactions on Mathematical Software,
830  ! Volume 5, Number 3, September 1979, pages 308-323.
831  !
832  ! Parameters:
833  !
834  ! Input, integer ( kind = 4 ) N, the number of entries in the vectors.
835  !
836  ! Input/output, real ( kind = 8 ) X(*), one of the vectors to swap.
837  !
838  ! Input, integer ( kind = 4 ) INCX, the increment between successive
839  ! entries of X.
840  !
841  ! Input/output, real ( kind = 8 ) Y(*), one of the vectors to swap.
842  !
843  ! Input, integer ( kind = 4 ) INCY, the increment between successive
844  ! elements of Y.
845  !
846  implicit none
847 
848  integer ( kind = 4 ) i
849  integer ( kind = 4 ) incx
850  integer ( kind = 4 ) incy
851  integer ( kind = 4 ) ix
852  integer ( kind = 4 ) iy
853  integer ( kind = 4 ) m
854  integer ( kind = 4 ) n
855  real ( kind = 8 ) temp
856  real ( kind = 8 ) x(*)
857  real ( kind = 8 ) y(*)
858 
859  if ( n <= 0 ) then
860 
861  else if ( incx == 1 .and. incy == 1 ) then
862 
863  m = mod( n, 3 )
864 
865  do i = 1, m
866  temp = x(i)
867  x(i) = y(i)
868  y(i) = temp
869  end do
870 
871  do i = m + 1, n, 3
872 
873  temp = x(i)
874  x(i) = y(i)
875  y(i) = temp
876 
877  temp = x(i+1)
878  x(i+1) = y(i+1)
879  y(i+1) = temp
880 
881  temp = x(i+2)
882  x(i+2) = y(i+2)
883  y(i+2) = temp
884 
885  end do
886 
887  else
888 
889  if ( 0 <= incx ) then
890  ix = 1
891  else
892  ix = ( - n + 1 ) * incx + 1
893  end if
894 
895  if ( 0 <= incy ) then
896  iy = 1
897  else
898  iy = ( - n + 1 ) * incy + 1
899  end if
900 
901  do i = 1, n
902  temp = x(ix)
903  x(ix) = y(iy)
904  y(iy) = temp
905  ix = ix + incx
906  iy = iy + incy
907  end do
908 
909  end if
910 
911  return
912  end
913  function idamax ( n, dx, incx )
914 
915  !*****************************************************************************80
916  !
917  !! IDAMAX indexes the array element of maximum absolute value.
918  !
919  ! Discussion:
920  !
921  ! This routine uses double precision real arithmetic.
922  !
923  ! Licensing:
924  !
925  ! This code is distributed under the GNU LGPL license.
926  !
927  ! Modified:
928  !
929  ! 08 April 1999
930  !
931  ! Author:
932  !
933  ! Original FORTRAN77 version by Charles Lawson, Richard Hanson,
934  ! David Kincaid, Fred Krogh.
935  ! FORTRAN90 version by John Burkardt.
936  !
937  ! Reference:
938  !
939  ! Jack Dongarra, Jim Bunch, Cleve Moler, Pete Stewart,
940  ! LINPACK User's Guide,
941  ! SIAM, 1979,
942  ! ISBN13: 978-0-898711-72-1,
943  ! LC: QA214.L56.
944  !
945  ! Charles Lawson, Richard Hanson, David Kincaid, Fred Krogh,
946  ! Algorithm 539,
947  ! Basic Linear Algebra Subprograms for Fortran Usage,
948  ! ACM Transactions on Mathematical Software,
949  ! Volume 5, Number 3, September 1979, pages 308-323.
950  !
951  ! Parameters:
952  !
953  ! Input, integer ( kind = 4 ) N, the number of entries in the vector.
954  !
955  ! Input, real ( kind = 8 ) X(*), the vector to be examined.
956  !
957  ! Input, integer ( kind = 4 ) INCX, the increment between successive
958  ! entries of SX.
959  !
960  ! Output, integer ( kind = 4 ) IDAMAX, the index of the element of SX of
961  ! maximum absolute value.
962  !
963  implicit none
964 
965  real ( kind = 8 ) dmax
966  real ( kind = 8 ) dx(*)
967  integer ( kind = 4 ) i
968  integer ( kind = 4 ) idamax
969  integer ( kind = 4 ) incx
970  integer ( kind = 4 ) ix
971  integer ( kind = 4 ) n
972 
973  idamax = 0
974 
975  if ( n < 1 .or. incx <= 0 ) then
976  return
977  end if
978 
979  idamax = 1
980 
981  if ( n == 1 ) then
982  return
983  end if
984 
985  if ( incx == 1 ) then
986 
987  dmax = abs( dx(1) )
988 
989  do i = 2, n
990  if ( dmax < abs( dx(i) ) ) then
991  idamax = i
992  dmax = abs( dx(i) )
993  end if
994  end do
995 
996  else
997 
998  ix = 1
999  dmax = abs( dx(1) )
1000  ix = ix + incx
1001 
1002  do i = 2, n
1003  if ( dmax < abs( dx(ix) ) ) then
1004  idamax = i
1005  dmax = abs( dx(ix) )
1006  end if
1007  ix = ix + incx
1008  end do
1009 
1010  end if
1011 
1012  return
1013  end
1014 
real(kind=8) function ddot(n, dx, incx, dy, incy)
Definition: blas1_d.f90:296
real(kind=8) function dnrm2(n, x, incx)
Definition: blas1_d.f90:388
integer(kind=4) function idamax(n, dx, incx)
Definition: blas1_d.f90:914
subroutine dscal(n, sa, x, incx)
Definition: blas1_d.f90:703
subroutine dswap(n, x, incx, y, incy)
Definition: blas1_d.f90:795
subroutine drot(n, x, incx, y, incy, c, s)
Definition: blas1_d.f90:482
subroutine daxpy(n, da, dx, incx, dy, incy)
Definition: blas1_d.f90:62
real(kind=8) function dasum(n, x, incx)
Definition: blas1_d.f90:2
subroutine drotg(sa, sb, c, s)
Definition: blas1_d.f90:587
subroutine dcopy(n, dx, incx, dy, incy)
Definition: blas1_d.f90:182