On Wed, Dec 6, 2017 at 6:52 PM, bvw1987 <address@hidden> wrote:
I have built a rather basic nonlinear optimization problem and have strived
to speed up the analysis by running it on a Linux machine to use
parallelization of the code. In so doing, I have noticed that when I ran the
exact same code on both sides except for the use of parallel functions that
I got different answers. One converged (Windows version) and the other
didn't converge.
The tolerances have been specified as the same for each including the
internal integrator, which for these purposes was 'lsode'. It is interesting
to note that the Windows Octave version is 4.2.1 and Linux is 4.0.0. The
Linux version has to load "odepkg" to use lsode while the Windows version
has lsode integral to it.
Do you all have any idea what is going on?
I would start by putting 4.2.1 on both systems.
--
Sent from: http://octave.1599824.n4.nabble.com/Octave-General- f1599825.html
_______________________________________________
Help-octave mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/help-octave