[Cfp-interest 2010] Re: AIs related to range errors clarifications

[Jim Thomas]

> On May 25, 2021, at 10:00 AM, Fred J. Tydeman <tydeman at tybor.com> wrote:
> 
> On Sat, 22 May 2021 21:08:44 -0700 Jim Thomas wrote:
>> 
>> The IEC 60559 model for correct rounding includes a (hypothetical) preliminary result correct to infinite range and precision. In C (outside of Annex F), which doesn't have this model and doesn't require correct rounding, what does "result with infinite precision" refer to?
> 
> I was thinking of result before final rounding.
> 
> I want to exclude zero and +/-min normal if they are exact from being underflow.
> 

Here’s a draft that excludes zero and +/- min normal. It also avoids referring to a hypothetical result (with infinite precision or ordinary accuracy).

The result underflows if it is nonzero and has magnitude (absolute value) less than the minimum normalized number in the type, or if it is inexact and has magnitude no greater than the minimum normalized number in the type.

- Jim Thomas

> 
> ---
> Fred J. Tydeman        Tydeman Consulting
> tydeman at tybor.com      Testing, numerics, programming
> +1 (702) 608-6093      Vice-chair of PL22.11 (ANSI "C")
> Sample C99+FPCE tests: http://www.tybor.com
> Savers sleep well, investors eat well, spenders work forever.
> 
> _______________________________________________
> Cfp-interest mailing list
> Cfp-interest at oakapple.net
> http://mailman.oakapple.net/mailman/listinfo/cfp-interest

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20210528/9e440bd4/attachment.htm>