This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: When is stage 1 scheduled to close for 3.4?
> Op do 01-05-2003, om 03:41 schreef Diego Novillo:
> > The work being done on the tree-ssa branch is reaching a state in
> > which we could start thinking about a merge. We're nearly done
> > with basic infrastructure work. We are now starting to look at
> > performance tuning.
> >
> > Depending on what the schedule looks like, it may make sense for
> > us to try merging the branch into mainline in time for 3.4.
It would be nice IMO, assuming that you can finish the merge in
foreseable future.
>
> That would be really cool, if feasable. A few questions:
>
> The SPEC scores so far don't show much runtime performance improvements
> and compile time is still worse, too (I have also seen compiler
> performance PRs that are fixed in tree-ssa because of GIMPLE, but...)
I would see they major point for merging tree-SSA at this point as a
validation of scheme. There are number of projects dependeing on
tree-SSA to be finished and these will get a push when they can go to
the mainline eventually.
> Do you have some target numbers for tree-ssa compile time performance
> and memory usage when all infrastructure work is finished (ie. un-SSA
> and copyprop working), and how much improvement do you expect from
> performance tuning.
> Any idea if it's possible to turn off one or two RTL optimizers ;-)?
>
> And wouldn't it be better to merge tree-ssa after the new register
> allocator is finished?
These two are independent long term projects, there is no reason to
block one because of the other.
Honza
>
> Greetz
> Steven
>