Re: elimination methods...

Michael Wolfe <mwolfe@cse.ogi.edu>
Fri, 23 Jun 1995 04:02:16 GMT

          From comp.compilers

Related articles
elimination methods... sreedhar@flo.cs.mcgill.ca (V.C. SREEDHAR) (1995-05-28)
Re: elimination methods... mwolfe@cse.ogi.edu (Michael Wolfe) (1995-06-23)
Re: elimination methods... csmith@convex.convex.com (1995-06-30)
Re: elimination methods... thomasl@groucho.csd.uu.se (1995-07-05)
| List of all articles for this month |

Newsgroups: comp.compilers
From: Michael Wolfe <mwolfe@cse.ogi.edu>
Keywords: optimize, courses
Organization: Oregon Graduate Institute - Computer Science & Engineering
References: 95-05-138
Date: Fri, 23 Jun 1995 04:02:16 GMT
Status: RO



I have taught compiler optimization courses as short courses several times;
the audience is almost entirely commercial compiler vendors, often for
processors that we don't usually consider (signal processing, etc.). I always
ask questions like what data flow analysis methods are used. The sampling is
not scientific, but there are always quite a few who use elimination methods
(interval analysis) and even some who use syntax-based methods, in addition
to those who use iterative methods. I ask the interval or syntax-based group
what they do in case of irreducible graphs, and the answers are either fall
back to iterative methods, or use node splitting, or give up.


no science here.
--
- Michael Wolfe
--


Post a followup to this message

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