[tree-ssa] Alias analysis speedups
Daniel Berlin
dberlin@dberlin.org
Fri Feb 21 20:27:00 GMT 2003
On Friday, February 21, 2003, at 03:03 PM, Toon Moene wrote:
> law@redhat.com wrote:
>
>> In message <1045844562.8671.18.camel@shadowfax>, Diego Novillo writes:
>
>> >Mixed blessing in this case. Better alias info may lead to longer
>> >compile times and memory problems. Hence our PHI pruning
>> heuristics.
>
>> Yes, but those are very extreme cases.
>
> "Elk nadeel hep se foordeel" :-)
>
> What would this mean for an extremely non-aliasing language like
> Fortran ?
>
Couldn't tell you experimentally if he wanted to, since Fortran doesn't
create GIMPLE trees.
But, in general, it wouldn't be a problem for an extreme non-aliasing
language, since there would generally be *no* alias sets (because of no
aliases) the sets, not a large number of *small* sets, as you end up
with in C.
Even if you had some aliases, it would still be a small number of sets.
In C, we have a large number of things that alias other things, but a
small number of things they each alias.
In fortran, you'd have a small number of things that alias other
things, and a small number of things they each alias.
> Inquiring minds want to know ...
>
> --
> Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
> Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
> Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
> GNU Fortran 95: http://gcc-g95.sourceforge.net/ (under construction)
>
More information about the Gcc-patches
mailing list