Re: Vector with >2^32 elements in 64 bit octave

[c.]

On 9 Nov 2011, at 11:02, Johan Lans wrote:

> Hi
>  I want to make a calculation with a vector larger than 2^32 elements.
> To be able to do this I've gotten access to a 64bit computer with a lot of 
> RAM,
> and installed octave 3.0.1. However, when I do 
> 
> octave:27> n = [1:2^31];
> 
> I get:
> 
> error: invalid range
> error: evaluating assignment expression near line 27, column 3
> 
> Why is this? I have "Octave was configured for "x86_64-pc-linux-gnu".",
> and there is no problem doing x=int64(2^40), for example.
> I'm not sure if I can install a newer version of octave, since I don't have 
> root access.
> 
> /Johan

To do that you have to use a copy of Octave built with the option:

--enable-64             (EXPERIMENTAL) use 64-bit integers for array
                          dimensions and indexing

Which seems to be off in the version you are using.

If you cannot convince your system administrator to install a binary built with 
that option 
you will have to compile Octave yourself, this can be done even without root 
access.

I wou
HTH,
c.