Re: high resolution floating point

[CdeMills]

Jordi Gutiérrez Hermoso-2 wrote
> On Thu, 2014-02-06 at 00:04 -0800, CdeMills wrote:
>> The real question is: do we intend to support Ginac on the long run
>> ?
> 
> I'm not a huge fan of GiNaC. I would like to see a different symbolic
> engine being used, such as Sage.

OK. For now, I renamed things as 'mpv' in the multi-precision toolbox; I
would like to fix them first and see later how we can interact with some
symbolic engine. A 'mpv' is an octave_value with a scalar, a prec
(precision) and a rnd (rounding).

Initialisation now works. The next point is that binary ops like arg1 op
arg2, to take full advantage of multi-precision, should re-formulate arg1
and arg2 with the maximum resolution of arg1 and arg2. This is to say that I
have to write a CAST function examining both args. Any example of such
approach ?

Regards

Pascal




--
View this message in context: 
http://octave.1599824.n4.nabble.com/high-resolution-floating-point-tp4661592p4661746.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.