WANTED: Rating of common optimization routines.

I wonder if there's any published report on the commonly employed
optimization routines/procedures targeted for scalar-type RISC based
applications such as:


o register allocation
o copy propagation
o common subexpression optimization
o peephole optimization
o strength reduction
o code motion of loop invariants
o ...


that is, regarding each one's contribution to the percentage of
improvement on the target execution speed when turned on. It would be
interesting to know which ones provides the highest improvement so
that emphasis can be put on them. I tend to think in general that
register allocation rates the highest, followed by peephole
optimization and strength reduction. Would appreciate any help on
getting the article(s).


Napi
--
http://www.cs.indiana.edu/hyplan/napi.html


--

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