Re: [boost] [metamath] Plans for compile time double

From: Peder Holt (peder.holt_at_[hidden])
Date: 2005-08-31 04:25:50

On 8/30/05, Cromwell Enage <sponage_at_[hidden]> wrote:
> --- Chris Weed wrote:
> > Hi,
> > I would like to help out with this. Was the plan to
> > implement most of the trigonometric functions with
> > taylor expansions or some other algorithm?
> > Chris
> 
> The plan was to implement the primary templates
> (sine_impl::apply, etc.) using Taylor expansions, then
> let the tag-dependent specializations apply their own
> algorithms if need be.  If you know what that other
> algorithm is, feel free to apply it to
> boost::metamath::double_.

It is basically the same algorithm, but since the basic operations on
double_ (times,plus,minus,divides) are designed to yield the same
result as for the runtime equivalent, we need to use
times_double_double etc, which uses a 61 bit mantissa, and then
truncate (round) to a 52 bit mantissa.
I guess we need to implement sine_double, cosine_double etc in order
to give the user a chance to exploit the entire 61 bit mantissa.

Regards,
Peder

> 
>                               Cromwell D. Enage
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
>