Re: arbitrary precision support opinions

[David Bateman]

Martin McDermott wrote:
> Hello
> 
> I wondering what people think about  arbitrary precision support(mostly
> for large integers).
> 
> I'm pretty new to Octave, but it does sound pretty handy and like it
> would put it a step above Matlab. Ive also been looking for a
> interesting project for awhile now, so I'm not expecting/demanding this
> feature out of anyone.
> 
> So, do you guys think it would be useful? Would it be used by anyone
> else? Any potential problems or general objections? 
> 
> At first it could just be a simple ifndef to choose, and maybe later it
> could be a option during runtime (I'm not sure how yet).
> 
> This would be a pretty big project for me so Id just like some input
> about it first.
> 
> Thanks
> Marty
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Help-octave mailing list
> address@hidden
> https://www.cae.wisc.edu/mailman/listinfo/help-octave


Matlab's fixed point toolbox is based on GMP and does arbitrary
precision integer math, but is painfully slow.. The octave fixed point
toolbox does arbitrary precision integer math with 1 to 31 bits and is
faster than the matlab fixed point code. This stuff is generally for
hardware implementations and so you typically want 8 to 16 bits in real
life.

D.