Re: Divergence from Solving functional nonlinear equation

[Jaroslav Hajek]

On Tue, Dec 29, 2009 at 4:43 PM, John W. Eaton <address@hidden> wrote:
> On 29-Dec-2009, Jason Martin wrote:
>
> |      Thanks!  I didn't know that little trick.  I will still recompile
> | for my windows due to the fact that I am running on a 64-bit system and
> | the version I have installed at the moment is for the 32-bit system of
> | Windows.  That is my main driving force for recompiling on my system.
>
> I don't think that will matter much either as Octave is by default a
> 32-bit application.
>
> Were you expecting to get something different if you build it for
> yourself on your 64-bit hardware?  That won't happen unless you
> use the --enable-64 option for Octave's configure script.
>
> It is not trivial to build Octave to take advantage of 64-bit indexing
> and allow manipulation of large arrays, primarily because many of the
> dependencies also need special treatment to handle 64-bit indexing in
> a way that is compatible with what Octave expects when using --enable-64.
>
> Also, there is no real benefit to building Octave this way unless you
> really plan to create large arrays (> 2GB) and you have a large amount
> of memory.  For example, if you plan to do more than a few trivial
> operations with a 2GB array, you will probably need more than 8GB of
> memory because Octave generates temporary copies for many operations.

I'd also like to note that the limit (in theory, at least) really is
about 2e9 elements, not 2GB data. In case of double arrays, 2e9
elements amounts to 16GB of data. I perfectly agree with the rest of
the comment.

hth

-- 
RNDr. Jaroslav Hajek, PhD
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz