Related articles |
---|
rearranging code invalidates liveness info fjh@munta.cs.mu.OZ.AU (1994-06-10) |
Re: rearranging code invalidates liveness info preston@noel.cs.rice.edu (1994-06-13) |
Re: rearranging code invalidates liveness info cliffc@noel.cs.rice.edu (1994-06-13) |
rearranging code invalidates liveness info ssimmons@convex.com (1994-06-13) |
Re: rearranging code invalidates liveness info newburn@aslan.ece.cmu.edu (1994-06-13) |
Re: rearranging code invalidates liveness info mernst@research.microsoft.com (1994-06-15) |
Re: rearranging code invalidates liveness info hbaker@netcom.com (1994-06-16) |
Re: rearranging code invalidates liveness info cliffc@noel.cs.rice.edu (1994-06-16) |
Re: rearranging code invalidates liveness info daniel@quilty.stanford.edu (1994-06-21) |
Newsgroups: | comp.compilers |
From: | ssimmons@convex.com (Steve Simmons) |
Keywords: | optimize |
Organization: | CONVEX News Network, Engineering (cnn.eng), Richardson, Tx USA |
References: | 94-06-070 |
Date: | Mon, 13 Jun 1994 16:44:26 GMT |
fjh@munta.cs.mu.OZ.AU (Fergus Henderson) writes:
> I'm part of a team developing a compiler (which happens to be for a logic
> programming language). We do quite a bit of analysis of the source code
> before generating code. The code is annotated with the information
> derived from this analysis, including liveness info, variable binding
> info, variable scope info, etc.
Hmmm... This is common problem... Anytime that you transform your
immediate representation, all static data structures must be
recalculated... Therefore, liveness values may need to be reset if hoist
your stores into another basic block. In fact, it is rarely possible to
keep your entire IR intact without having to do a recalculation of
specific data structures that other phases may be dependent upon.
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.