Re: Loop optimizer issues

[Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>]

Hello,

> How about this:
> 
>      1. We identify generic bits of infrastructure that are
>         IL-independent and merge those into tree-ssa.
>      2. RTL optimizations are either converted to use the tree-ssa
>         framework or deemed useful only for RTL, in which case I guess
>         we wouldn't lose anything by merging them.

sounds fine to me.  I have just started trying to make the pieces I need
to work on tree-ssa level, so in about a week I should be able to say
with reasonable confidence which parts are really of this type.

> However, I would like to see these bits in independent chunks, if
> possible.  One alternative that I've been thinking about is to create a
> sub-branch of tree-ssa to deal exclusively with loop optimizations and
> vectorization:
> 
>      1. The tree-ssa branch would like to become 3.5 in the near
>         future.  It already contains a lot of experimental code that I
>         would very much like to stabilize.  Adding even more
>         experimental code at this stage seems a bit risky to me.
>      2. Since work for loop optimizers and vectorization has just
>         started, it is unclear how much of it will be ready for 3.5.
>      3. A sub-branch of tree-ssa is no different than a mainline
>         branch.  Code that's ready for 3.5, can be merged at the same
>         time as tree-ssa.
> 
> Thoughts?

Also seems like a reasonable idea to me.

Zdenek