Newsgroups: comp.arch
Path: utzoo!henry
From: henry@zoo.toronto.edu (Henry Spencer)
Subject: Re: IEEE arithmetic
Message-ID: <1991Jun3.222604.21057@zoo.toronto.edu>
Date: Mon, 3 Jun 1991 22:26:04 GMT
References: <9106032144.AA15477@ucbvax.Berkeley.EDU>
Organization: U of Toronto Zoology

In article <9106032144.AA15477@ucbvax.Berkeley.EDU> jbs@watson.ibm.com writes:
>>Having had some need to compute hypotenuses without overflow in code that
>>otherwise wanted to use single precision for speed, I'm afraid I disagree.
>
>... In deciding whether a feature is of value one must
>look at alternatives.  In your example the obvious alternative is to
>compute the hypotenuses using double precision...

Right, that is, using extended single precision.  That's why the extended
precisions are there, and why 754 very strongly recommends that an extended
precision be available for any normal precision.  What if I'd been using
double precision already?
-- 
"We're thinking about upgrading from    | Henry Spencer @ U of Toronto Zoology
SunOS 4.1.1 to SunOS 3.5."              |  henry@zoo.toronto.edu  utzoo!henry