Re: Please listen to Ulrich here...
On 2013-08-26 18:13:46 +0200, Ulrich Kulisch wrote:
> Am 24.08.2013 22:30, schrieb Vincent Lefevre:
> >On 2013-08-24 12:49:46 -0700, G. William (Bill) Walster wrote:
> >>As far as I can tell the only time when a case can be made that EDP is
> >>essential for interval computations is when all interval inputs are
^^^^^^^^^^^^^^^^^^^^^
> >>degenerate and therefore infinitely precise.
> >Even in this case (and you want the tightest accuracy mode), EDP is
> >not essential. You just need a correctly rounded dot product.
> >
> Vincent,
>
> a correctly rounded dot product suffices for many applications. But
> there are many other applications where an EDP is needed.
[...]
My point is that EDP is useless (compared to correctly-rounded dot
product) in interval arithmetic specification if you don't have an
*interval type* that can represent the exact result.
It might be useful in some implementations, just like any feature
one may want (e.g. exact/accurate polynomial evaluation, useful for
math functions, or iRRAM-like behavior), but that's just internals,
and how operations are implemented is out of the scope of the
standard.
--
Vincent Lefèvre <vincent@xxxxxxxxxx> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)