Re: Faster divides

cdg@nullstone.com (Christopher Glaeser)
Wed, 25 Oct 1995 21:05:29 GMT

          From comp.compilers

Related articles
Faster divides sdenman@wolf.cs.washington.edu (1995-10-14)
Re: Faster divides idr@cs.pdx.edu (1995-10-23)
Re: Faster divides cdg@nullstone.com (1995-10-25)
| List of all articles for this month |

Newsgroups: comp.compilers
From: cdg@nullstone.com (Christopher Glaeser)
Keywords: arithmetic, performance
Organization: Compilers Central
References: 95-10-120
Date: Wed, 25 Oct 1995 21:05:29 GMT

> shifted by 16 won't fit in 32 bits. Which could be a problem.


When implementing faster divides, it is important to consider these and
other potential problems. Recently, I evaluated the implementations of
integer division for twenty compilers on UNIX, PC, and Mac platforms.
Twelve of these compilers generated incorrect code for integer division.


The most common defect was related to power-of-two division optimization,
but other problems included compile-time core dumps for some bit-patterns,
and exponential optimization algorithms that required several hours to
compile a single divide expression for other bit-patterns.


Christopher Glaeser cdg@nullstone.com
Nullstone Corporation
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.