On Tuesday, March 27, 2012 5:20:55 AM UTC+5:30, rl wrote:
>
> Hi,
>
>
> > This also shows that we should definitely support using mpmath to do
> > numerics, not just numpy.
>
> The question is for speed, I assume numpy to be much
> faster than mpmath. Maybe one should support both
> as "backends" for the interval arithmetic.
>
Yeah I think we should support both for interval arithmetic. Numpy
can be the default option while mpmath is used when specified 
explicitly. 

> > Figure 13 is an equation that enumerates
> > all possible j x 17 bitmaps.  The one for k =
> > 
> 9960939379918958884971672962127852754715004339660129306651505519271702802395266424689642842174350718121267153782770623355993237280874144307891
> > 
> 325963941337723487857735749823926629715517173716995165232890538221612403238855866184013235585136048828693337902491454229288667081096184496091
> > 
> 705183454067827731551705405381627380967602565625016981482083418783163849115590225610003652351370343874461848378737238198224849863465033159410
> > 
> 054974700593138339226497249461751545728366702369745461014655997933798537483143786841806593422227898388722980000748404719
> > just happens to be one that looks like the equation itself.
>
> A nice example, too :-)
>
> >  Clearly we need a high precision library like mpmath to do this.
>
> And we need a very good interval arithmetic library.
> What is available inside of mpmath? What is missing?
> How to implement the missing parts?
>
mpmath has a very standard interval arithmetic library, but we need
property checking for domain, continuity for Algorithm 3.2-3.4.
So I think I have to extend the classes in mpmath. Jeff Tupper's
(the author of the paper on graphing implicit equations) M.Sc thesis 
was on interval arithmetic, and he has explained a lot about implementing
these properties.(http://www.dgp.utoronto.ca/people/mooncake/msc.html).

I have not gone through it completely, but I think it is detailed enough to 
have an implementation. I have exams going on, so I am unable to read 
through it completely. I will look into it in a couple of days and include 
it
in my gsoc proposal.
Thanks,
Bharath M R

> All these questions are obviously part of the project
> although they do not deal with plotting directly.
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sympy/-/gPSZD5gUXgEJ.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/sympy?hl=en.

Reply via email to