[Octave-bug-tracker] [bug #31731] Test failed for sqp

[Marco Caliari]

Follow-up Comment #2, bug #31731 (project octave):

I'm on a linux system, i686.
With the value x_opt found in the test, I see


phi(x_opt) = 5.39498477702741e-02
max (g(x_opt)) = 3.55271367880050e-15


while with the result x I get (with tol = eps) 


phi(x) = 5.39498477702744e-02
max (g(x_opt)) = 5.32907051820075e-15


(so, both x_opt and x are double precision accurate), but


norm (x_opt-x, inf) = 1.80132234683938e-08


The result I get is satisfactory for me, but I cannot understand the
difference.

Marco

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?31731>

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