Divide-by-multiply algorithm reference needed

"Joseph M. Newcomer" <newcomer@flounder.com>
Mon, 13 Sep 2010 01:03:18 -0400

          From comp.compilers
Related articles
Divide-by-multiply algorithm reference needed newcomer@flounder.com (Joseph M. Newcomer) (2010-09-13)
Re: Divide-by-multiply algorithm reference needed rsc@swtch.com (Russ Cox) (2010-09-13)
Re: Divide-by-multiply algorithm reference needed gah@ugcs.caltech.edu (glen herrmannsfeldt) (2010-09-14)
Re: Divide-by-multiply algorithm reference needed newcomer@flounder.com (Joseph M. Newcomer) (2010-09-14)
Re: Divide-by-multiply algorithm reference needed newcomer@flounder.com (Joseph M. Newcomer) (2010-09-14)
Re: Divide-by-multiply algorithm reference needed pvk@pvk.ca (Paul Khuong) (2010-09-14)
Re: Divide-by-multiply algorithm reference needed gah@ugcs.caltech.edu (glen herrmannsfeldt) (2010-09-15)
| List of all articles for this month |
From: "Joseph M. Newcomer" <newcomer@flounder.com>
Newsgroups: comp.compilers
Date: Mon, 13 Sep 2010 01:03:18 -0400
Organization: NewsGuy - Unlimited Usenet $19.95
Keywords: arithmetic, question
Posted-Date: 13 Sep 2010 21:02:58 EDT

Division can be approximated with multiprecision multiplication, e.g.,
32-bit division can be done by producing the appropriate 64-bit
product and using the high order 32 bits.


My problem is I am trying to find the algorithm by which I can
generate the 32-bit multiplier and the shift amount, given a specific
constant divisor. I am unable to find any references via google. Any
recommendations? No, I really don't want to read the gcc source (even
if it does this) unless someone knows the specific subroutine name I
should be looking at.


joe
Joseph M. Newcomer [MVP]
email: newcomer@flounder.com
Web: http://www.flounder.com
MVP Tips: http://www.flounder.com/mvp_tips.htm


Post a followup to this message

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