c\BeginDoc
c
c\Name: pdseupd
c
c Message Passing Layer: MPI
c
c\Description:
c
c This subroutine returns the converged approximations to eigenvalues
c of A*z = lambda*B*z and (optionally):
c
c (1) the corresponding approximate eigenvectors,
c
c (2) an orthonormal (Lanczos) basis for the associated approximate
c invariant subspace,
c
c (3) Both.
c
c There is negligible additional cost to obtain eigenvectors. An orthonormal
c (Lanczos) basis is always computed. There is an additional storage cost
c of n*nev if both are requested (in this case a separate array Z must be
c supplied).
c
c These quantities are obtained from the Lanczos factorization computed
c by PSSAUPD for the linear operator OP prescribed by the MODE selection
c (see IPARAM(7) in PSSAUPD documentation.) PSSAUPD must be called before
c this routine is called. These approximate eigenvalues and vectors are
c commonly called Ritz values and Ritz vectors respectively. They are
c referred to as such in the comments that follow. The computed orthonormal
c basis for the invariant subspace corresponding to these Ritz values is
c referred to as a Lanczos basis.
c
c See documentation in the header of the subroutine PSSAUPD for a definition
c of OP as well as other terms and the relation of computed Ritz values
c and vectors of OP with respect to the given problem A*z = lambda*B*z.
c
c The approximate eigenvalues of the original problem are returned in
c ascending algebraic order. The user may elect to call this routine
c once for each desired Ritz vector and store it peripherally if desired.
c There is also the option of computing a selected set of these vectors
c with a single call.
c
c\Usage:
c call pdseupd
c ( COMM, RVEC, HOWMNY, SELECT, D, Z, LDZ, SIGMA, BMAT, N, WHICH, NEV, TOL,
c RESID, NCV, V, LDV, IPARAM, IPNTR, WORKD, WORKL, LWORKL, INFO )
c
c\Arguments
c COMM MPI Communicator for the processor grid. (INPUT)
c
c RVEC LOGICAL (INPUT)
c Specifies whether Ritz vectors corresponding to the Ritz value
c approximations to the eigenproblem A*z = lambda*B*z are computed.
c
c RVEC = .FALSE. Compute Ritz values only.
c
c RVEC = .TRUE. Compute Ritz vectors.
c
c HOWMNY Character*1 (INPUT)
c Specifies how many Ritz vectors are wanted and the form of Z
c the matrix of Ritz vectors. See remark 1 below.
c = 'A': compute NEV Ritz vectors;
c = 'S': compute some of the Ritz vectors, specified
c by the logical array SELECT.
c
c SELECT Logical array of dimension NCV. (INPUT/WORKSPACE)
c If HOWMNY = 'S', SELECT specifies the Ritz vectors to be
c computed. To select the Ritz vector corresponding to a
c Ritz value D(j), SELECT(j) must be set to .TRUE..
c If HOWMNY = 'A' , SELECT is used as workspace.
c
c D Double precision array of dimension NEV. (OUTPUT)
c On exit, D contains the Ritz value approximations to the
c eigenvalues of A*z = lambda*B*z. The values are returned
c in ascending order. If IPARAM(7) = 3,4,5 then D represents
c the Ritz values of OP computed by pdsaupd transformed to
c those of the original eigensystem A*z = lambda*B*z. If
c IPARAM(7) = 1,2 then the Ritz values of OP are the same
c as the those of A*z = lambda*B*z.
c
c Z Double precision N by NEV array if HOWMNY = 'A'. (OUTPUT)
c On exit, Z contains the B-orthonormal Ritz vectors of the
c eigensystem A*z = lambda*B*z corresponding to the Ritz
c value approximations.
c If RVEC = .FALSE. then Z is not referenced.
c NOTE: The array Z may be set equal to first NEV columns of the
c Arnoldi/Lanczos basis array V computed by PSSAUPD.
c
c LDZ Integer. (INPUT)
c The leading dimension of the array Z. If Ritz vectors are
c desired, then LDZ .ge. max( 1, N ). In any case, LDZ .ge. 1.
c
c SIGMA Double precision (INPUT)
c If IPARAM(7) = 3,4,5 represents the shift. Not referenced if
c IPARAM(7) = 1 or 2.
c
c
c **** The remaining arguments MUST be the same as for the ****
c **** call to PDNAUPD that was just completed. ****
c
c NOTE: The remaining arguments
c
c BMAT, N, WHICH, NEV, TOL, RESID, NCV, V, LDV, IPARAM, IPNTR,
c WORKD, WORKL, LWORKL, INFO
c
c must be passed directly to PSSEUPD following the last call
c to PSSAUPD. These arguments MUST NOT BE MODIFIED between
c the the last call to PSSAUPD and the call to PSSEUPD.
c
c Two of these parameters (WORKL, INFO) are also output parameters:
c
c WORKL Double precision work array of length LWORKL. (OUTPUT/WORKSPACE)
c WORKL(1:4*ncv) contains information obtained in
c PSSAUPD. They are not changed by PSSEUPD.
c WORKL(4*ncv+1:ncv*ncv+8*ncv) holds the
c untransformed Ritz values, the computed error estimates,
c and the associated eigenvector matrix of H.
c
c Note: IPNTR(8:10) contains the pointers into WORKL for addresses
c of the above information computed by PSSEUPD.
c -------------------------------------------------------------
c IPNTR(8): pointer to the NCV RITZ values of the original system.
c IPNTR(9): pointer to the NCV corresponding error bounds.
c IPNTR(10): pointer to the NCV by NCV matrix of eigenvectors
c of the tridiagonal matrix T. Only referenced by
c PSSEUPD if RVEC = .TRUE. See Remarks.
c -------------------------------------------------------------
c
c INFO Integer. (OUTPUT)
c Error flag on output.
c = 0: Normal exit.
c = -1: N must be positive.
c = -2: NEV must be positive.
c = -3: NCV must be greater than NEV and less than or equal to N.
c = -5: WHICH must be one of 'LM', 'SM', 'LA', 'SA' or 'BE'.
c = -6: BMAT must be one of 'I' or 'G'.
c = -7: Length of private work WORKL array is not sufficient.
c = -8: Error return from trid. eigenvalue calculation;
c Information error from LAPACK routine dsteqr .
c = -9: Starting vector is zero.
c = -10: IPARAM(7) must be 1,2,3,4,5.
c = -11: IPARAM(7) = 1 and BMAT = 'G' are incompatible.
c = -12: NEV and WHICH = 'BE' are incompatible.
c = -14: PSSAUPD did not find any eigenvalues to sufficient
c accuracy.
c = -15: HOWMNY must be one of 'A' or 'S' if RVEC = .true.
c = -16: HOWMNY = 'S' not yet implemented
c = -17: DSEUPD got a different count of the number of converged
c Ritz values than DSAUPD got. This indicates the user
c probably made an error in passing data from DSAUPD to
c DSEUPD or that the data was modified before entering
c DSEUPD .
c
c\BeginLib
c
c\References:
c 1. D.C. Sorensen, "Implicit Application of Polynomial Filters in
c a k-Step Arnoldi Method", SIAM J. Matr. Anal. Apps., 13 (1992),
c pp 357-385.
c 2. R.B. Lehoucq, "Analysis and Implementation of an Implicitly
c Restarted Arnoldi Iteration", Rice University Technical Report
c TR95-13, Department of Computational and Applied Mathematics.
c 3. B.N. Parlett, "The Symmetric Eigenvalue Problem". Prentice-Hall,
c 1980.
c 4. B.N. Parlett, B. Nour-Omid, "Towards a Black Box Lanczos Program",
c Computer Physics Communications, 53 (1989), pp 169-179.
c 5. B. Nour-Omid, B.N. Parlett, T. Ericson, P.S. Jensen, "How to
c Implement the Spectral Transformation", Math. Comp., 48 (1987),
c pp 663-673.
c 6. R.G. Grimes, J.G. Lewis and H.D. Simon, "A Shifted Block Lanczos
c Algorithm for Solving Sparse Symmetric Generalized Eigenproblems",
c SIAM J. Matr. Anal. Apps., January (1993).
c 7. L. Reichel, W.B. Gragg, "Algorithm 686: FORTRAN Subroutines
c for Updating the QR decomposition", ACM TOMS, December 1990,
c Volume 16 Number 4, pp 369-377.
c
c\Remarks
c 1. The converged Ritz values are always returned in increasing
c (algebraic) order.
c
c 2. Currently only HOWMNY = 'A' is implemented. It is included at this
c stage for the user who wants to incorporate it.
c
c\Routines called:
c dsesrt ARPACK routine that sorts an array X, and applies the
c corresponding permutation to a matrix A.
c dsortr dsortr ARPACK sorting routine.
c pdnorm2 Parallel ARPACK routine that computes the 2-norm of a vector.
c pivout Parallel ARPACK utility routine that prints integers.
c pdvout Parallel ARPACK utility routine that prints vectors.
c dgeqr2 LAPACK routine that computes the QR factorization of
c a matrix.
c dlacpy LAPACK matrix copy routine.
c pdlamch ScaLAPACK routine that determines machine constants.
c dorm2r LAPACK routine that applies an orthogonal matrix in
c factored form.
c dsteqr LAPACK routine that computes eigenvalues and eigenvectors
c of a tridiagonal matrix.
c dger Level 2 BLAS rank one update to a matrix.
c dcopy Level 1 BLAS that copies one vector to another .
c dscal Level 1 BLAS that scales a vector.
c dswap Level 1 BLAS that swaps the contents of two vectors.
c\Authors
c Danny Sorensen Phuong Vu
c Richard Lehoucq CRPC / Rice University
c Chao Yang Houston, Texas
c Dept. of Computational &
c Applied Mathematics
c Rice University
c Houston, Texas
c
c\Parallel Modifications
c Kristi Maschhoff
c
c\Revision history:
c Starting Point: Serial Code FILE: seupd.F SID: 2.4
c
c\SCCS Information:
c FILE: seupd.F SID: 1.10 DATE OF SID: 04/10/01
c
c\EndLib
c
c-----------------------------------------------------------------------
subroutine pdseupd
& (comm , rvec , howmny, select, d ,
& z , ldz , sigma , bmat , n ,
& which , nev , tol , resid , ncv ,
& v , ldv , iparam, ipntr , workd,
& workl , lworkl, info )
c
c %--------------------%
c | MPI Communicator |
c %--------------------%
c
integer comm
c
c %----------------------------------------------------%
c | Include files for debugging and timing information |
c %----------------------------------------------------%
c
include 'debug.h'
include 'stat.h'
c
c %------------------%
c | Scalar Arguments |
c %------------------%
c
character bmat, howmny, which*2
logical rvec
integer info, ldz, ldv, lworkl, n, ncv, nev
Double precision
& sigma, tol
c
c %-----------------%
c | Array Arguments |
c %-----------------%
c
integer iparam(7), ipntr(11)
logical select(ncv)
Double precision
& d(nev), resid(n), v(ldv,ncv), z(ldz, nev),
& workd(2*n), workl(lworkl)
c
c %------------%
c | Parameters |
c %------------%
c
Double precision
& one, zero
parameter (one = 1.0 , zero = 0.0 )
c
c %---------------%
c | Local Scalars |
c %---------------%
c
character type*6
integer bounds , ierr , ih , ihb , ihd ,
& iq , iw , j , k , ldh ,
& ldq , mode , msglvl, nconv , next ,
& ritz , irz , ibd , np , ishift,
& leftptr, rghtptr, numcnv, jj
Double precision
& bnorm2, rnorm, temp, temp1, eps23
logical reord
c
c %----------------------%
c | External Subroutines |
c %----------------------%
c
external dcopy , dger , dgeqr2 , dlacpy , dorm2r , dscal ,
& dsesrt , dsteqr , dswap , pdvout , pivout, dsortr
c
c %--------------------%
c | External Functions |
c %--------------------%
c
Double precision
& pdnorm2 , pdlamch10
external pdnorm2 , pdlamch10
c
c %---------------------%
c | Intrinsic Functions |
c %---------------------%
c
intrinsic min
c
c %-----------------------%
c | Executable Statements |
c %-----------------------%
c
c %------------------------%
c | Set default parameters |
c %------------------------%
c
msglvl = mseupd
mode = iparam(7)
nconv = iparam(5)
info = 0
c
c %--------------%
c | Quick return |
c %--------------%
c
if (nconv .eq. 0) go to 9000
ierr = 0
c
if (nconv .le. 0) ierr = -14
if (n .le. 0) ierr = -1
if (nev .le. 0) ierr = -2
if (ncv .le. nev) ierr = -3
if (which .ne. 'LM' .and.
& which .ne. 'SM' .and.
& which .ne. 'LA' .and.
& which .ne. 'SA' .and.
& which .ne. 'BE') ierr = -5
if (bmat .ne. 'I' .and. bmat .ne. 'G') ierr = -6
if ( (howmny .ne. 'A' .and.
& howmny .ne. 'P' .and.
& howmny .ne. 'S') .and. rvec )
& ierr = -15
if (rvec .and. howmny .eq. 'S') ierr = -16
c
if (rvec .and. lworkl .lt. ncv**2+8*ncv) ierr = -7
c
if (mode .eq. 1 .or. mode .eq. 2) then
type = 'REGULR'
else if (mode .eq. 3 ) then
type = 'SHIFTI'
else if (mode .eq. 4 ) then
type = 'BUCKLE'
else if (mode .eq. 5 ) then
type = 'CAYLEY'
else
ierr = -10
end if
if (mode .eq. 1 .and. bmat .eq. 'G') ierr = -11
if (nev .eq. 1 .and. which .eq. 'BE') ierr = -12
c
c %------------%
c | Error Exit |
c %------------%
c
if (ierr .ne. 0) then
info = ierr
go to 9000
end if
c
c %-------------------------------------------------------%
c | Pointer into WORKL for address of H, RITZ, BOUNDS, Q |
c | etc... and the remaining workspace. |
c | Also update pointer to be used on output. |
c | Memory is laid out as follows: |
c | workl(1:2*ncv) := generated tridiagonal matrix H |
c | The subdiagonal is stored in workl(2:ncv). |
c | The dead spot is workl(1) but upon exiting |
c | pdsaupd stores the B-norm of the last residual |
c | vector in workl(1). We use this
c | workl(2*ncv+1:2*ncv+ncv) := ritz values |
c | The wanted values are in the first NCONV spots. |
c | workl(3*ncv+1:3*ncv+ncv) := computed Ritz estimates |
c | The wanted values are in the first NCONV spots. |
c | NOTE: workl(1:4*ncv) is set by pdsaupd and is not |
c | modified by pdseupd . |
c %-------------------------------------------------------%
c
c %-------------------------------------------------------%
c | The following is used and set by pdseupd . |
c | workl(4*ncv+1:4*ncv+ncv) := used as workspace during |
c | computation of the eigenvectors of H. Stores |
c | the diagonal of H. Upon EXIT contains the NCV |
c | Ritz values of the original system. The first |
c | NCONV spots have the wanted values. If MODE = |
c | 1 or 2 then will equal workl(2*ncv+1:3*ncv). |
c | workl(5*ncv+1:5*ncv+ncv) := used as workspace during |
c | computation of the eigenvectors of H. Stores |
c | the subdiagonal of H. Upon EXIT contains the |
c | NCV corresponding Ritz estimates of the |
c | original system. The first NCONV spots have the |
c | wanted values. If MODE = 1,2 then will equal |
c | workl(3*ncv+1:4*ncv). |
c | workl(6*ncv+1:6*ncv+ncv*ncv) := orthogonal Q that is |
c | the eigenvector matrix for H as returned by |
c | dsteqr . Not referenced if RVEC = .False. |
c | Ordering follows that of workl(4*ncv+1:5*ncv) |
c | workl(6*ncv+ncv*ncv+1:6*ncv+ncv*ncv+2*ncv) := |
c | Workspace. Needed by dsteqr and by pdseupd . |
c | GRAND total of NCV*(NCV+8) locations. |
c %-------------------------------------------------------%
c
c
ih = ipntr(5)
ritz = ipntr(6)
bounds = ipntr(7)
ldh = ncv
ldq = ncv
ihd = bounds + ldh
ihb = ihd + ldh
iq = ihb + ldh
iw = iq + ldh*ncv
next = iw + 2*ncv
ipntr(4) = next
ipntr(8) = ihd
ipntr(9) = ihb
ipntr(10) = iq
c
c %----------------------------------------%
c | irz points to the Ritz values computed |
c | by _seigt before exiting _saup2. |
c | ibd points to the Ritz estimates |
c | computed by _seigt before exiting |
c | _saup2. |
c %----------------------------------------%
c
irz = ipntr(11)+ncv
ibd = irz+ncv
c
c
c %---------------------------------%
c | Set machine dependent constant. |
c %---------------------------------%
c
eps23 = pdlamch10(comm, 'Epsilon-Machine')
eps23 = eps23**(2.0 / 3.0 )
c
c %---------------------------------------%
c | RNORM is B-norm of the RESID(1:N). |
c | BNORM2 is the 2 norm of B*RESID(1:N). |
c | Upon exit of pdsaupd WORKD(1:N) has |
c | B*RESID(1:N). |
c %---------------------------------------%
c
rnorm = workl(ih)
if (bmat .eq. 'I') then
bnorm2 = rnorm
else if (bmat .eq. 'G') then
bnorm2 = pdnorm2 (comm, n, workd, 1)
end if
c
if (msglvl .gt. 2) then
call pdvout (comm, logfil, ncv, workl(irz), ndigit,
& '_seupd: Ritz values passed in from _SAUPD.')
call pdvout (comm, logfil, ncv, workl(ibd), ndigit,
& '_seupd: Ritz estimates passed in from _SAUPD.')
end if
if (rvec) then
c
reord = .false.
c
c %---------------------------------------------------%
c | Use the temporary bounds array to store indices |
c | These will be used to mark the select array later |
c %---------------------------------------------------%
c
do 10 j = 1,ncv
workl(bounds+j-1) = j
select(j) = .false.
10 continue
c
c %-------------------------------------%
c | Select the wanted Ritz values. |
c | Sort the Ritz values so that the |
c | wanted ones appear at the tailing |
c | NEV positions of workl(irz). Move |
c | the corresponding error estimates |
c | in workl(bound) accordingly. |
c %-------------------------------------%
c
np = ncv - nev
ishift = 0
call pdsgets (comm , ishift, which ,
& nev , np , workl(irz),
& workl(bounds), workl , workl(np+1))
c
if (msglvl .gt. 2) then
call pdvout (comm, logfil, ncv, workl(irz), ndigit,
& '_seupd: Ritz values after calling _SGETS.')
call pdvout (comm, logfil, ncv, workl(bounds), ndigit,
& '_seupd: Ritz value indices after calling _SGETS.')
end if
c
c %-----------------------------------------------------%
c | Record indices of the converged wanted Ritz values |
c | Mark the select array for possible reordering |
c %-----------------------------------------------------%
c
numcnv = 0
do 11 j = 1,ncv
temp1 = max(eps23, abs(workl(irz+ncv-j)) )
jj = workl(bounds + ncv - j)
if (numcnv .lt. nconv .and.
& workl(ibd+jj-1) .le. tol*temp1) then
select(jj) = .true.
numcnv = numcnv + 1
if (jj .gt. nev) reord = .true.
endif
11 continue
c
c %-----------------------------------------------------------%
c | Check the count (numcnv) of converged Ritz values with |
c | the number (nconv) reported by _saupd. If these two |
c | are different then there has probably been an error |
c | caused by incorrect passing of the _saupd data. |
c %-----------------------------------------------------------%
c
if (msglvl .gt. 2) then
call pivout(comm, logfil, 1, [numcnv], ndigit,
& '_neupd: Number of specified eigenvalues')
call pivout(comm, logfil, 1, [nconv], ndigit,
& '_neupd: Number of "converged" eigenvalues')
end if
c
if (numcnv .ne. nconv) then
info = -17
go to 9000
end if
c
c %-----------------------------------------------------------%
c | Call LAPACK routine _steqr to compute the eigenvalues and |
c | eigenvectors of the final symmetric tridiagonal matrix H. |
c | Initialize the eigenvector matrix Q to the identity. |
c %-----------------------------------------------------------%
c
call dcopy (ncv-1, workl(ih+1) , 1, workl(ihb), 1)
call dcopy (ncv , workl(ih+ldh), 1, workl(ihd), 1)
c
call dsteqr ('Identity', ncv , workl(ihd),
& workl(ihb), workl(iq), ldq ,
& workl(iw) , ierr)
c
if (ierr .ne. 0) then
info = -8
go to 9000
end if
c
if (msglvl .gt. 1) then
call dcopy (ncv, workl(iq+ncv-1), ldq, workl(iw), 1)
call pdvout (comm, logfil, ncv, workl(ihd), ndigit,
& '_seupd: NCV Ritz values of the final H matrix')
call pdvout (comm, logfil, ncv, workl(iw), ndigit,
& '_seupd: last row of the eigenvector matrix for H')
end if
c
if (reord) then
c
c %---------------------------------------------%
c | Reordered the eigenvalues and eigenvectors |
c | computed by _steqr so that the "converged" |
c | eigenvalues appear in the first NCONV |
c | positions of workl(ihd), and the associated |
c | eigenvectors appear in the first NCONV |
c | columns. |
c %---------------------------------------------%
c
leftptr = 1
rghtptr = ncv
c
if (ncv .eq. 1) go to 30
c
20 if (select(leftptr)) then
c
c %-------------------------------------------%
c | Search, from the left, for the first Ritz |
c | value that has not converged. |
c %-------------------------------------------%
c
leftptr = leftptr + 1
c
else if (.not.select(rghtptr)) then
c
c %----------------------------------------------%
c | Search, from the right, the first Ritz value |
c | that has converged. |
c %----------------------------------------------%
c
rghtptr = rghtptr - 1
c
else
c
c %----------------------------------------------%
c | Swap the Ritz value on the left that has not |
c | converged with the Ritz value on the right |
c | that has converged. Swap the associated |
c | eigenvector of the tridiagonal matrix H as |
c | well. |
c %----------------------------------------------%
c
temp = workl(ihd+leftptr-1)
workl(ihd+leftptr-1) = workl(ihd+rghtptr-1)
workl(ihd+rghtptr-1) = temp
call dcopy (ncv, workl(iq+ncv*(leftptr-1)), 1,
& workl(iw), 1)
call dcopy (ncv, workl(iq+ncv*(rghtptr-1)), 1,
& workl(iq+ncv*(leftptr-1)), 1)
call dcopy (ncv, workl(iw), 1,
& workl(iq+ncv*(rghtptr-1)), 1)
leftptr = leftptr + 1
rghtptr = rghtptr - 1
c
end if
c
if (leftptr .lt. rghtptr) go to 20
c
30 end if
c
if (msglvl .gt. 2) then
call pdvout (comm, logfil, ncv, workl(ihd), ndigit,
& '_seupd: The eigenvalues of H--reordered')
end if
c
c %----------------------------------------%
c | Load the converged Ritz values into D. |
c %----------------------------------------%
c
call dcopy (nconv, workl(ihd), 1, d, 1)
c
else
c
c %-----------------------------------------------------%
c | Ritz vectors not required. Load Ritz values into D. |
c %-----------------------------------------------------%
c
call dcopy (nconv, workl(ritz), 1, d, 1)
call dcopy (ncv, workl(ritz), 1, workl(ihd), 1)
c
end if
c
c %------------------------------------------------------------------%
c | Transform the Ritz values and possibly vectors and corresponding |
c | Ritz estimates of OP to those of A*x=lambda*B*x. The Ritz values |
c | (and corresponding data) are returned in ascending order. |
c %------------------------------------------------------------------%
c
if (type .eq. 'REGULR') then
c
c %---------------------------------------------------------%
c | Ascending sort of wanted Ritz values, vectors and error |
c | bounds. Not necessary if only Ritz values are desired. |
c %---------------------------------------------------------%
c
if (rvec) then
call dsesrt ('LA', rvec , nconv, d, ncv, workl(iq), ldq)
else
call dcopy (ncv, workl(bounds), 1, workl(ihb), 1)
end if
c
else
c
c %-------------------------------------------------------------%
c | * Make a copy of all the Ritz values. |
c | * Transform the Ritz values back to the original system. |
c | For TYPE = 'SHIFTI' the transformation is |
c | lambda = 1/theta + sigma |
c | For TYPE = 'BUCKLE' the transformation is |
c | lambda = sigma * theta / ( theta - 1 ) |
c | For TYPE = 'CAYLEY' the transformation is |
c | lambda = sigma * (theta + 1) / (theta - 1 ) |
c | where the theta are the Ritz values returned by pdsaupd . |
c | NOTES: |
c | *The Ritz vectors are not affected by the transformation. |
c | They are only reordered. |
c %-------------------------------------------------------------%
c
call dcopy (ncv, workl(ihd), 1, workl(iw), 1)
if (type .eq. 'SHIFTI') then
do 40 k=1, ncv
workl(ihd+k-1) = one / workl(ihd+k-1) + sigma
40 continue
else if (type .eq. 'BUCKLE') then
do 50 k=1, ncv
workl(ihd+k-1) = sigma * workl(ihd+k-1) /
& (workl(ihd+k-1) - one)
50 continue
else if (type .eq. 'CAYLEY') then
do 60 k=1, ncv
workl(ihd+k-1) = sigma * (workl(ihd+k-1) + one) /
& (workl(ihd+k-1) - one)
60 continue
end if
c
c %-------------------------------------------------------------%
c | * Store the wanted NCONV lambda values into D. |
c | * Sort the NCONV wanted lambda in WORKL(IHD:IHD+NCONV-1) |
c | into ascending order and apply sort to the NCONV theta |
c | values in the transformed system. We will need this to |
c | compute Ritz estimates in the original system. |
c | * Finally sort the lambda`s into ascending order and apply |
c | to Ritz vectors if wanted. Else just sort lambda`s into |
c | ascending order. |
c | NOTES: |
c | *workl(iw:iw+ncv-1) contain the theta ordered so that they |
c | match the ordering of the lambda. We`ll use them again for |
c | Ritz vector purification. |
c %-------------------------------------------------------------%
c
call dcopy (nconv, workl(ihd), 1, d, 1)
call dsortr ('LA', .true., nconv, workl(ihd), workl(iw))
if (rvec) then
call dsesrt ('LA', rvec , nconv, d, ncv, workl(iq), ldq)
else
call dcopy (ncv, workl(bounds), 1, workl(ihb), 1)
call dscal (ncv, bnorm2/rnorm, workl(ihb), 1)
call dsortr ('LA', .true., nconv, d, workl(ihb))
end if
c
end if
c
c %------------------------------------------------%
c | Compute the Ritz vectors. Transform the wanted |
c | eigenvectors of the symmetric tridiagonal H by |
c | the Lanczos basis matrix V. |
c %------------------------------------------------%
c
if (rvec .and. howmny .eq. 'A') then
c
c %----------------------------------------------------------%
c | Compute the QR factorization of the matrix representing |
c | the wanted invariant subspace located in the first NCONV |
c | columns of workl(iq,ldq). |
c %----------------------------------------------------------%
c
call dgeqr2 (ncv, nconv , workl(iq) ,
& ldq, workl(iw+ncv), workl(ihb),
& ierr)
c
c %--------------------------------------------------------%
c | * Postmultiply V by Q. |
c | * Copy the first NCONV columns of VQ into Z. |
c | The N by NCONV matrix Z is now a matrix representation |
c | of the approximate invariant subspace associated with |
c | the Ritz values in workl(ihd). |
c %--------------------------------------------------------%
c
call dorm2r ('Right' , 'Notranspose', n ,
& ncv , nconv , workl(iq),
& ldq , workl(iw+ncv), v ,
& ldv , workd(n+1) , ierr )
call dlacpy ('All', n, nconv, v, ldv, z, ldz)
c
c %-----------------------------------------------------%
c | In order to compute the Ritz estimates for the Ritz |
c | values in both systems, need the last row of the |
c | eigenvector matrix. Remember, it`s in factored form |
c %-----------------------------------------------------%
c
do 65 j = 1, ncv-1
workl(ihb+j-1) = zero
65 continue
workl(ihb+ncv-1) = one
call dorm2r ('Left', 'Transpose' , ncv ,
& 1 , nconv , workl(iq) ,
& ldq , workl(iw+ncv), workl(ihb),
& ncv , temp , ierr )
c
else if (rvec .and. howmny .eq. 'S') then
c
c Not yet implemented. See remark 2 above.
c
end if
c
if (type .eq. 'REGULR' .and. rvec) then
c
do 70 j=1, ncv
workl(ihb+j-1) = rnorm * abs( workl(ihb+j-1) )
70 continue
c
else if (type .ne. 'REGULR' .and. rvec) then
c
c %-------------------------------------------------%
c | * Determine Ritz estimates of the theta. |
c | If RVEC = .true. then compute Ritz estimates |
c | of the theta. |
c | If RVEC = .false. then copy Ritz estimates |
c | as computed by pdsaupd . |
c | * Determine Ritz estimates of the lambda. |
c %-------------------------------------------------%
c
call dscal (ncv, bnorm2, workl(ihb), 1)
if (type .eq. 'SHIFTI') then
c
do 80 k=1, ncv
workl(ihb+k-1) = abs( workl(ihb+k-1) )
& / workl(iw+k-1)**2
80 continue
c
else if (type .eq. 'BUCKLE') then
c
do 90 k=1, ncv
workl(ihb+k-1) = sigma * abs( workl(ihb+k-1) )
& / ( workl(iw+k-1)-one )**2
90 continue
c
else if (type .eq. 'CAYLEY') then
c
do 100 k=1, ncv
workl(ihb+k-1) = abs( workl(ihb+k-1)
& / workl(iw+k-1)*(workl(iw+k-1)-one) )
100 continue
c
end if
c
end if
c
if (type .ne. 'REGULR' .and. msglvl .gt. 1) then
call pdvout (comm, logfil, nconv, d, ndigit,
& '_seupd: Untransformed converged Ritz values')
call pdvout (comm, logfil, nconv, workl(ihb), ndigit,
& '_seupd: Ritz estimates of the untransformed Ritz values')
else if (msglvl .gt. 1) then
call pdvout (comm, logfil, nconv, d, ndigit,
& '_seupd: Converged Ritz values')
call pdvout (comm, logfil, nconv, workl(ihb), ndigit,
& '_seupd: Associated Ritz estimates')
end if
c
c %-------------------------------------------------%
c | Ritz vector purification step. Formally perform |
c | one of inverse subspace iteration. Only used |
c | for MODE = 3,4,5. See reference 7 |
c %-------------------------------------------------%
c
if (rvec .and. (type .eq. 'SHIFTI' .or. type .eq. 'CAYLEY')) then
c
do 110 k=0, nconv-1
workl(iw+k) = workl(iq+k*ldq+ncv-1)
& / workl(iw+k)
110 continue
c
else if (rvec .and. type .eq. 'BUCKLE') then
c
do 120 k=0, nconv-1
workl(iw+k) = workl(iq+k*ldq+ncv-1)
& / (workl(iw+k)-one)
120 continue
c
end if
c
if (type .ne. 'REGULR')
& call dger (n, nconv, one, resid, 1, workl(iw), 1, z, ldz)
c
9000 continue
c
return
c
c %----------------%
c | End of pdseupd |
c %----------------%
c
end
