Block Scheduling: determ vs nondeterm

PAULLONG@delphi.com
Sun, 5 Jun 1994 01:56:05 GMT

          From comp.compilers

Related articles
Block Scheduling: determ vs nondeterm PAULLONG@delphi.com (1994-06-05)
Re: Block Scheduling: determ vs nondeterm preston@noel.cs.rice.edu (1994-06-13)
Re: Block Scheduling: determ vs nondeterm paullong@delphi.com (1994-06-15)
Re: Block Scheduling: determ vs nondeterm preston@noel.cs.rice.edu (1994-06-16)
Re: Block Scheduling: determ vs nondeterm chase@Think.COM (1994-06-16)
Re: Block Scheduling: determ vs nondeterm grunwald@foobar.cs.colorado.edu (1994-06-20)
| List of all articles for this month |
Newsgroups: comp.compilers
From: PAULLONG@delphi.com
Keywords: optimize, question
Organization: Compilers Central
Date: Sun, 5 Jun 1994 01:56:05 GMT

In order to remove as many jumps from one basic block to another, I
understand that one needs to perform block scheduling, much like
instruction scheduling. However, I've never written a block scheduler and
don't know what the traditional approach is. My first inclination is to
write a non-deterministic block scheduler using a genetic algorithm to
obtain a soltuion that is "good enough." Are traditional block schedulers
deterministic? If so, what is the basic approach? Have others written
non-deterministic block schedulers? If so, again what was the basic
approach?


paullong@delphi.com
--


Post a followup to this message

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