Re: rearranging code invalidates liveness info

fjh@munta.cs.mu.OZ.AU (Fergus Henderson) writes:
>The problem I've run into is that this organization seems to make code
>reorganization for optimization purposes quite difficult, since anything
>but the most trivial rearrangements will invalidate the annotations on the
>code. I'm sure this problem must have been encountered by many of you
>before. How do compiler writers usually handle this problem?


People writing experimental compilers tend to redo analyses after any
significant transformation. It's easy, correct, and not usually too
slow. People writing production compilers have a harder time, hence
the interest on incremental data-flow analysis. An example of an
industrial approach is described in


    author="Suneel Jain and Carol Thompson",
    title="An Efficient Approach to Data Flow Analysis in a Multi Pass
                                  Global Optimizer",
    pages="154--163",
    journal=sigplan,
    year=1988,
    month=jul,
    volume=23,
    number=7,
    note=pldi88


Preston Briggs
--

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