Re: Type theory vs floating-point arithmetic

• To: Joe Darcy <Joe.Darcy@Sun.COM>
• Subject: Re: Type theory vs floating-point arithmetic
• From: Ross Casley <rcasley@cisco.com>
• Date: Wed, 03 Sep 2003 16:27:46 -0700
• Cc: Achim Jung <A.Jung@cs.bham.ac.uk>, types@cis.upenn.edu, Matthias Blume <blume@tti-c.org>
• In-Reply-To: <200309031307.h83D7iKt027056@saul.cis.upenn.edu>
• References: <Pine.GSO.4.56.0309021131350.27288@preston.cs.bham.ac.uk> <200308311405.h7VE5WTV011241@saul.cis.upenn.edu> <Pine.GSO.4.56.0309021131350.27288@preston.cs.bham.ac.uk>

At 09:14 PM 9/2/2003 -0700, Joe Darcy wrote:
>For example, the value
>"3.0" can be written as
>
>  0x1.8p1
>
>that is "1.8" as a hex string (i.e. 1.5 decimal) times 2 to the 1.  The
>double floating-point value next larger than 3.0 is
>
>  0x1.800000000001p1
>
>and the number next smaller is
>
>  0x1.7999 9999 9999p1

Isn't it 0x1.7FFFFFFFFFFFp1 ?

-Ross

• Re: Type theory vs floating-point arithmetic
• From: Joe Darcy <Joe.Darcy@Sun.COM>

• Re: Type theory vs floating-point arithmetic
• From: Achim Jung <A.Jung@cs.bham.ac.uk>

• Prev: Re: Type theory vs floating-point arithmetic
• Next: Re: Type theory vs floating-point arithmetic
• Index(es):
  • Main
  • Thread