Related articles |
---|
optimization and register allocation sastry@GODEL.MIEL.MOT.COM (1995-02-24) |
Re: optimization and register allocation davidm@Rational.COM (1995-02-28) |
Re: optimization and register allocation sethml@puree.ugcs.caltech.edu (1995-03-04) |
Re: optimization and register allocation preston@tera.com (1995-03-06) |
Newsgroups: | comp.compilers |
From: | sethml@puree.ugcs.caltech.edu (Seth LaForge) |
Keywords: | optimize, registers, comment |
Organization: | California Institute of Technology, Pasadena, CA |
References: | 95-02-182 95-03-009 |
Date: | Sat, 4 Mar 1995 12:32:31 GMT |
David Moore <davidm@Rational.COM> wrote:
>Note that f2-a is not quite the same as (-a)+f2 because the
>latter may overflow while the former does not. In C you
How is this? Suppose a == -200 and f2 == 100, and you are evaluating
in an 8-bit context... It seems to me that both would overflow...
Seth
[-a can overflow on a two's complement machine if it's the largest negative
number, while in that case f2-a would not overflow if f2 is any negative
number. -John]
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.