LAPACK  3.6.1 LAPACK: Linear Algebra PACKage
 subroutine cpprfs ( character UPLO, integer N, integer NRHS, complex, dimension( * ) AP, complex, dimension( * ) AFP, complex, dimension( ldb, * ) B, integer LDB, complex, dimension( ldx, * ) X, integer LDX, real, dimension( * ) FERR, real, dimension( * ) BERR, complex, dimension( * ) WORK, real, dimension( * ) RWORK, integer INFO )

CPPRFS

Purpose:
``` CPPRFS improves the computed solution to a system of linear
equations when the coefficient matrix is Hermitian positive definite
and packed, and provides error bounds and backward error estimates
for the solution.```
Parameters
 [in] UPLO ``` UPLO is CHARACTER*1 = 'U': Upper triangle of A is stored; = 'L': Lower triangle of A is stored.``` [in] N ``` N is INTEGER The order of the matrix A. N >= 0.``` [in] NRHS ``` NRHS is INTEGER The number of right hand sides, i.e., the number of columns of the matrices B and X. NRHS >= 0.``` [in] AP ``` AP is COMPLEX array, dimension (N*(N+1)/2) The upper or lower triangle of the Hermitian matrix A, packed columnwise in a linear array. The j-th column of A is stored in the array AP as follows: if UPLO = 'U', AP(i + (j-1)*j/2) = A(i,j) for 1<=i<=j; if UPLO = 'L', AP(i + (j-1)*(2n-j)/2) = A(i,j) for j<=i<=n.``` [in] AFP ``` AFP is COMPLEX array, dimension (N*(N+1)/2) The triangular factor U or L from the Cholesky factorization A = U**H*U or A = L*L**H, as computed by SPPTRF/CPPTRF, packed columnwise in a linear array in the same format as A (see AP).``` [in] B ``` B is COMPLEX array, dimension (LDB,NRHS) The right hand side matrix B.``` [in] LDB ``` LDB is INTEGER The leading dimension of the array B. LDB >= max(1,N).``` [in,out] X ``` X is COMPLEX array, dimension (LDX,NRHS) On entry, the solution matrix X, as computed by CPPTRS. On exit, the improved solution matrix X.``` [in] LDX ``` LDX is INTEGER The leading dimension of the array X. LDX >= max(1,N).``` [out] FERR ``` FERR is REAL array, dimension (NRHS) The estimated forward error bound for each solution vector X(j) (the j-th column of the solution matrix X). If XTRUE is the true solution corresponding to X(j), FERR(j) is an estimated upper bound for the magnitude of the largest element in (X(j) - XTRUE) divided by the magnitude of the largest element in X(j). The estimate is as reliable as the estimate for RCOND, and is almost always a slight overestimate of the true error.``` [out] BERR ``` BERR is REAL array, dimension (NRHS) The componentwise relative backward error of each solution vector X(j) (i.e., the smallest relative change in any element of A or B that makes X(j) an exact solution).``` [out] WORK ` WORK is COMPLEX array, dimension (2*N)` [out] RWORK ` RWORK is REAL array, dimension (N)` [out] INFO ``` INFO is INTEGER = 0: successful exit < 0: if INFO = -i, the i-th argument had an illegal value```
Internal Parameters:
`  ITMAX is the maximum number of steps of iterative refinement.`
Date
November 2011

Definition at line 173 of file cpprfs.f.

173 *
174 * -- LAPACK computational routine (version 3.4.0) --
175 * -- LAPACK is a software package provided by Univ. of Tennessee, --
176 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
177 * November 2011
178 *
179 * .. Scalar Arguments ..
180  CHARACTER uplo
181  INTEGER info, ldb, ldx, n, nrhs
182 * ..
183 * .. Array Arguments ..
184  REAL berr( * ), ferr( * ), rwork( * )
185  COMPLEX afp( * ), ap( * ), b( ldb, * ), work( * ),
186  \$ x( ldx, * )
187 * ..
188 *
189 * ====================================================================
190 *
191 * .. Parameters ..
192  INTEGER itmax
193  parameter ( itmax = 5 )
194  REAL zero
195  parameter ( zero = 0.0e+0 )
196  COMPLEX cone
197  parameter ( cone = ( 1.0e+0, 0.0e+0 ) )
198  REAL two
199  parameter ( two = 2.0e+0 )
200  REAL three
201  parameter ( three = 3.0e+0 )
202 * ..
203 * .. Local Scalars ..
204  LOGICAL upper
205  INTEGER count, i, ik, j, k, kase, kk, nz
206  REAL eps, lstres, s, safe1, safe2, safmin, xk
207  COMPLEX zdum
208 * ..
209 * .. Local Arrays ..
210  INTEGER isave( 3 )
211 * ..
212 * .. External Subroutines ..
213  EXTERNAL caxpy, ccopy, chpmv, clacn2, cpptrs, xerbla
214 * ..
215 * .. Intrinsic Functions ..
216  INTRINSIC abs, aimag, max, real
217 * ..
218 * .. External Functions ..
219  LOGICAL lsame
220  REAL slamch
221  EXTERNAL lsame, slamch
222 * ..
223 * .. Statement Functions ..
224  REAL cabs1
225 * ..
226 * .. Statement Function definitions ..
227  cabs1( zdum ) = abs( REAL( ZDUM ) ) + abs( aimag( zdum ) )
228 * ..
229 * .. Executable Statements ..
230 *
231 * Test the input parameters.
232 *
233  info = 0
234  upper = lsame( uplo, 'U' )
235  IF( .NOT.upper .AND. .NOT.lsame( uplo, 'L' ) ) THEN
236  info = -1
237  ELSE IF( n.LT.0 ) THEN
238  info = -2
239  ELSE IF( nrhs.LT.0 ) THEN
240  info = -3
241  ELSE IF( ldb.LT.max( 1, n ) ) THEN
242  info = -7
243  ELSE IF( ldx.LT.max( 1, n ) ) THEN
244  info = -9
245  END IF
246  IF( info.NE.0 ) THEN
247  CALL xerbla( 'CPPRFS', -info )
248  RETURN
249  END IF
250 *
251 * Quick return if possible
252 *
253  IF( n.EQ.0 .OR. nrhs.EQ.0 ) THEN
254  DO 10 j = 1, nrhs
255  ferr( j ) = zero
256  berr( j ) = zero
257  10 CONTINUE
258  RETURN
259  END IF
260 *
261 * NZ = maximum number of nonzero elements in each row of A, plus 1
262 *
263  nz = n + 1
264  eps = slamch( 'Epsilon' )
265  safmin = slamch( 'Safe minimum' )
266  safe1 = nz*safmin
267  safe2 = safe1 / eps
268 *
269 * Do for each right hand side
270 *
271  DO 140 j = 1, nrhs
272 *
273  count = 1
274  lstres = three
275  20 CONTINUE
276 *
277 * Loop until stopping criterion is satisfied.
278 *
279 * Compute residual R = B - A * X
280 *
281  CALL ccopy( n, b( 1, j ), 1, work, 1 )
282  CALL chpmv( uplo, n, -cone, ap, x( 1, j ), 1, cone, work, 1 )
283 *
284 * Compute componentwise relative backward error from formula
285 *
286 * max(i) ( abs(R(i)) / ( abs(A)*abs(X) + abs(B) )(i) )
287 *
288 * where abs(Z) is the componentwise absolute value of the matrix
289 * or vector Z. If the i-th component of the denominator is less
290 * than SAFE2, then SAFE1 is added to the i-th components of the
291 * numerator and denominator before dividing.
292 *
293  DO 30 i = 1, n
294  rwork( i ) = cabs1( b( i, j ) )
295  30 CONTINUE
296 *
297 * Compute abs(A)*abs(X) + abs(B).
298 *
299  kk = 1
300  IF( upper ) THEN
301  DO 50 k = 1, n
302  s = zero
303  xk = cabs1( x( k, j ) )
304  ik = kk
305  DO 40 i = 1, k - 1
306  rwork( i ) = rwork( i ) + cabs1( ap( ik ) )*xk
307  s = s + cabs1( ap( ik ) )*cabs1( x( i, j ) )
308  ik = ik + 1
309  40 CONTINUE
310  rwork( k ) = rwork( k ) + abs( REAL( AP( KK+K-1 ) ) )*
311  \$ xk + s
312  kk = kk + k
313  50 CONTINUE
314  ELSE
315  DO 70 k = 1, n
316  s = zero
317  xk = cabs1( x( k, j ) )
318  rwork( k ) = rwork( k ) + abs( REAL( AP( KK ) ) )*xk
319  ik = kk + 1
320  DO 60 i = k + 1, n
321  rwork( i ) = rwork( i ) + cabs1( ap( ik ) )*xk
322  s = s + cabs1( ap( ik ) )*cabs1( x( i, j ) )
323  ik = ik + 1
324  60 CONTINUE
325  rwork( k ) = rwork( k ) + s
326  kk = kk + ( n-k+1 )
327  70 CONTINUE
328  END IF
329  s = zero
330  DO 80 i = 1, n
331  IF( rwork( i ).GT.safe2 ) THEN
332  s = max( s, cabs1( work( i ) ) / rwork( i ) )
333  ELSE
334  s = max( s, ( cabs1( work( i ) )+safe1 ) /
335  \$ ( rwork( i )+safe1 ) )
336  END IF
337  80 CONTINUE
338  berr( j ) = s
339 *
340 * Test stopping criterion. Continue iterating if
341 * 1) The residual BERR(J) is larger than machine epsilon, and
342 * 2) BERR(J) decreased by at least a factor of 2 during the
343 * last iteration, and
344 * 3) At most ITMAX iterations tried.
345 *
346  IF( berr( j ).GT.eps .AND. two*berr( j ).LE.lstres .AND.
347  \$ count.LE.itmax ) THEN
348 *
349 * Update solution and try again.
350 *
351  CALL cpptrs( uplo, n, 1, afp, work, n, info )
352  CALL caxpy( n, cone, work, 1, x( 1, j ), 1 )
353  lstres = berr( j )
354  count = count + 1
355  GO TO 20
356  END IF
357 *
358 * Bound error from formula
359 *
360 * norm(X - XTRUE) / norm(X) .le. FERR =
361 * norm( abs(inv(A))*
362 * ( abs(R) + NZ*EPS*( abs(A)*abs(X)+abs(B) ))) / norm(X)
363 *
364 * where
365 * norm(Z) is the magnitude of the largest component of Z
366 * inv(A) is the inverse of A
367 * abs(Z) is the componentwise absolute value of the matrix or
368 * vector Z
369 * NZ is the maximum number of nonzeros in any row of A, plus 1
370 * EPS is machine epsilon
371 *
372 * The i-th component of abs(R)+NZ*EPS*(abs(A)*abs(X)+abs(B))
373 * is incremented by SAFE1 if the i-th component of
374 * abs(A)*abs(X) + abs(B) is less than SAFE2.
375 *
376 * Use CLACN2 to estimate the infinity-norm of the matrix
377 * inv(A) * diag(W),
378 * where W = abs(R) + NZ*EPS*( abs(A)*abs(X)+abs(B) )))
379 *
380  DO 90 i = 1, n
381  IF( rwork( i ).GT.safe2 ) THEN
382  rwork( i ) = cabs1( work( i ) ) + nz*eps*rwork( i )
383  ELSE
384  rwork( i ) = cabs1( work( i ) ) + nz*eps*rwork( i ) +
385  \$ safe1
386  END IF
387  90 CONTINUE
388 *
389  kase = 0
390  100 CONTINUE
391  CALL clacn2( n, work( n+1 ), work, ferr( j ), kase, isave )
392  IF( kase.NE.0 ) THEN
393  IF( kase.EQ.1 ) THEN
394 *
395 * Multiply by diag(W)*inv(A**H).
396 *
397  CALL cpptrs( uplo, n, 1, afp, work, n, info )
398  DO 110 i = 1, n
399  work( i ) = rwork( i )*work( i )
400  110 CONTINUE
401  ELSE IF( kase.EQ.2 ) THEN
402 *
403 * Multiply by inv(A)*diag(W).
404 *
405  DO 120 i = 1, n
406  work( i ) = rwork( i )*work( i )
407  120 CONTINUE
408  CALL cpptrs( uplo, n, 1, afp, work, n, info )
409  END IF
410  GO TO 100
411  END IF
412 *
413 * Normalize error.
414 *
415  lstres = zero
416  DO 130 i = 1, n
417  lstres = max( lstres, cabs1( x( i, j ) ) )
418  130 CONTINUE
419  IF( lstres.NE.zero )
420  \$ ferr( j ) = ferr( j ) / lstres
421 *
422  140 CONTINUE
423 *
424  RETURN
425 *
426 * End of CPPRFS
427 *
subroutine cpptrs(UPLO, N, NRHS, AP, B, LDB, INFO)
CPPTRS
Definition: cpptrs.f:110
subroutine chpmv(UPLO, N, ALPHA, AP, X, INCX, BETA, Y, INCY)
CHPMV
Definition: chpmv.f:151
subroutine xerbla(SRNAME, INFO)
XERBLA
Definition: xerbla.f:62
subroutine ccopy(N, CX, INCX, CY, INCY)
CCOPY
Definition: ccopy.f:52
real function slamch(CMACH)
SLAMCH
Definition: slamch.f:69
subroutine caxpy(N, CA, CX, INCX, CY, INCY)
CAXPY
Definition: caxpy.f:53
logical function lsame(CA, CB)
LSAME
Definition: lsame.f:55
subroutine clacn2(N, V, X, EST, KASE, ISAVE)
CLACN2 estimates the 1-norm of a square matrix, using reverse communication for evaluating matrix-vec...
Definition: clacn2.f:135

Here is the call graph for this function:

Here is the caller graph for this function: