[petsc-users] EPS monitor

[Kong, Fande]

Solver details:

EPS Object: 1 MPI processes
  type: jd
    search subspace is orthogonalized
    block size=1
    type of the initial subspace: non-Krylov
    size of the subspace after restarting: 6
    number of vectors after restarting from the previous iteration: 1
  problem type: generalized non-symmetric eigenvalue problem
  extraction type: harmonic Ritz
  selected portion of the spectrum: smallest eigenvalues in magnitude
  number of eigenvalues (nev): 1
  number of column vectors (ncv): 18
  maximum dimension of projected problem (mpd): 18
  maximum number of iterations: 10000
  tolerance: 0.0001
  convergence test: relative to the eigenvalue
BV Object: 1 MPI processes
  type: svec
  18 columns of global length 4225
  vector orthogonalization method: classical Gram-Schmidt
  orthogonalization refinement: if needed (eta: 0.7071)
  block orthogonalization method: Gram-Schmidt
  doing matmult as a single matrix-matrix product
DS Object: 1 MPI processes
  type: gnhep
ST Object: 1 MPI processes
  type: precond
  shift: 0.
  number of matrices: 2
  all matrices have different nonzero pattern
  KSP Object: (st_) 1 MPI processes
    type: bcgsl
      Ell = 2
      Delta = 0
    maximum iterations=10000, initial guess is zero
    tolerances:  relative=1e-05, absolute=1e-50, divergence=10000.
    left preconditioning
    using PRECONDITIONED norm type for convergence test
  PC Object: (st_) 1 MPI processes
    type: jacobi
    linear system matrix = precond matrix:
    Mat Object: () 1 MPI processes
      type: seqaij
      rows=4225, cols=4225
      total: nonzeros=37249, allocated nonzeros=37249
      total number of mallocs used during MatSetValues calls =0
        not using I-node routines

On Wed, Aug 16, 2017 at 2:12 PM, Jose E. Roman <jroman at dsic.upv.es> wrote:

>
> > El 16 ago 2017, a las 21:36, Kong, Fande <fande.kong at inl.gov> escribió:
> >
> > Hi All,
> >
> > How to understand the following messages:
> >
> >   1 EPS nconv=0 first unconverged value (error) 2.06312 (3.29164033e-01)
> >   2 EPS nconv=0 first unconverged value (error) 2.03951 (1.76223074e-01)
> >   3 EPS nconv=0 first unconverged value (error) 2.01177 (5.71109559e-02)
> >   4 EPS nconv=0 first unconverged value (error) 2.01042 (4.84609300e-02)
> >   5 EPS nconv=0 first unconverged value (error) 2.00708 (3.19917457e-02)
> >   6 EPS nconv=0 first unconverged value (error) 2.00595 (2.62792109e-02)
> >   7 EPS nconv=0 first unconverged value (error) 2.00504 (2.13766150e-02)
> >   8 EPS nconv=0 first unconverged value (error) 2.00441 (1.85066774e-02)
> >   9 EPS nconv=0 first unconverged value (error) 2.00397 (1.73188449e-02)
> >  10 EPS nconv=0 first unconverged value (error) 2.00366 (1.54528517e-02)
> >  11 EPS nconv=0 first unconverged value (error) 2.00339 (1.32215899e-02)
> >  12 EPS nconv=0 first unconverged value (error) 2.00316 (1.32215899e-02)
> >  13 EPS nconv=0 first unconverged value (error) 2.00316 (1.17928920e-02)
> >  14 EPS nconv=0 first unconverged value (error) 2.00297 (1.04964387e-02)
> >  15 EPS nconv=0 first unconverged value (error) 2.0028 (9.58244972e-03)
> >  16 EPS nconv=0 first unconverged value (error) 2.00268 (9.06634973e-03)
> >  17 EPS nconv=0 first unconverged value (error) 2.00198 (3.43444441e-04)
> >  18 EPS nconv=1 first unconverged value (error) 2.25718 (1.79769313e+308)
> >  18 EPS converged value (error) #0 2.00197 (5.69451918e-09)
> >
> >
> > When the solver converged, the wrong eigenvalue and the wrong residual
> are printed out. Do we design like this way?
> >
> > Fande,
>
> Is this the POWER solver? Most solvers in EPS approximate several
> eigenvalues simultaneously, but this is not the case in POWER - when one
> eigenvalue converges there is no approximation available for the next one.
>
> I will think about a simple fix.
>
> Jose
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20170816/d381b08c/attachment.html>