[Octave-bug-tracker] [bug #53700] eigs test failure related to ARPACK generating real NaN rather than complex NaN+1i*NaN

[John W. Eaton]

Follow-up Comment #29, bug #53700 (project octave):

After the patch on the stable branch, I see the following behavior"


octave:1> A = magic (100) / 10 + eye (100);
octave:2> opts.v0 = (1:100)';
octave:3> opts.maxit = 10;
octave:4> for i=9:20; d = eigs (A, i, "sm", opts); printf ("%d: %s", i,
disp(d(end))); end
warning: eigs: Only 2 of the 9 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
9:  NaN + NaNi
warning: eigs: Only 2 of the 10 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
10:  NaN + NaNi
warning: eigs: Only 4 of the 11 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
11:  NaN + NaNi
warning: eigs: Only 4 of the 12 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
12:  NaN + NaNi
warning: eigs: Only 4 of the 13 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
13:  NaN + NaNi
warning: eigs: Only 6 of the 14 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
14:  NaN + NaNi
warning: eigs: Only 7 of the 15 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
15:  NaN + NaNi
warning: eigs: Only 7 of the 16 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
16:  NaN + NaNi
warning: eigs: Only 9 of the 17 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
17:  NaN + NaNi
warning: eigs: Only 13 of the 18 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
18:  NaN + NaNi
warning: eigs: Only 11 of the 19 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
19:  NaN + NaNi
warning: eigs: Only 15 of the 20 requested eigenvalues converged
warning: called from
    eigs at line 285 column 18
20:  NaN + NaNi


I begin to see non-NaN results at i=49:


49:  1.0000e+00 - 9.9142e-13i
50:  1.0000e+00 + 4.9814e-13i





    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?53700>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/