[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Octave-bug-tracker] [bug #31731] Test failed for sqp
From: |
Rik |
Subject: |
[Octave-bug-tracker] [bug #31731] Test failed for sqp |
Date: |
Wed, 24 Nov 2010 03:32:22 +0000 |
User-agent: |
Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.12) Gecko/20101027 Ubuntu/9.10 (karmic) Firefox/3.6.12 |
Follow-up Comment #3, bug #31731 (project octave):
Strange. Your machine is obviously converging to a slightly different value.
For comparison, when I run the test I get
x =
-1.71714346231343e+00
1.59570956502671e+00
1.82724595374969e+00
-7.63643097800157e-01
-7.63643082600814e-01
and
phi(x)
ans = 5.39498477702777e-02
and finally
norm(x-x_opt,inf)
ans = 7.36525367361907e-08
It's interesting that the part of the test which is failing is
all (abs (x-x_opt) < 5*sqrt (eps))
but your result of 1.8e-8 is smaller than my result of 7e-8 and my test still
passes. Maybe it's worth splitting the test into two separate assert
statements to guarantee that nothing odd is happening.
%! assert (all (abs (x-x_opt) < 5*sqrt (eps)))
%! assert (abs (obj-obj_opt) < sqrt (eps))
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?31731>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/