Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: IEEEP1788

On 21.04.2015 22:12, Ralph Baker Kearfott wrote:

Ulrich et al,

On 04/21/2015 11:34 AM, Ulrich Kulisch wrote:
.
.
.

In contrast to the simplicity of arithmetics in IR and IF, IEEE P1788
develops interval arithmetic on the base of IEEE 754 arithmetic with all
its exceptions.


Does it?  How do you come to this conclusion?  The 754 data type
is featured as an option, but where do you see IEEE 754 exceptions
are used?


This is a big mistake. It unnecessarily pulls all the
IEEE 754 exceptions into interval arithmetic.


Please say where in the document 754 exceptions are used, rather than
just the data type.  How are the 754 exceptions interwoven into
the 1788 exceptions, namely, decorations.  Howe does it "pull all of the
754 exceptions into interval arithmetic" ???

.
.
.

By the way, IEEE 754-2008 will expire in 4 years unless a revision
PAR is submitted.  I am presently looking for people to lead the
effort to submit a PAR, revies 754-2008, and possibly revise it.

Baker
If I remember correctly, the use of IEEE 754 data types in a P1788 conforming implementation is optional and it is not required to provide/use any of the IEEE 754 exceptions.
If an implementation uses IEEE 754 data types, then P1788 has some requirements for how to use them, e.g. representation of [0, 0] by (-0.0, +0.0). However, P1788 does not discriminate between +0/-0 on the mathematic/arithmetic level. I find this very convenient, because IEEE 754 data types are commonly used and the standard document guides implementations in a common direction.
The only place, where IEEE 754 specific arithmetic is actually used in the standard document, is in chapter 12.12.12 (reduction operations on IEEE 754 data types), where arithmetic with NaNs and signed zeros is required. Again, this only applies if the implementation supports IEEE 754 data types. 

Oliver Heimlich