Re: [tree-ssa] AST optimizer in C++?

[Chris Lattner <sabre@nondot.org>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1000 bytes --]

On Sun, 25 Aug 2002, Diego Novillo wrote:
> On Sun, 25 Aug 2002, Pop Sébastian wrote:

> > Diego, maybe it is worth to include PHI nodes directly as a
> > SIMPLE extension?
> >
> I am not too keen on SSA re-writing schemes.  When you encode SSA
> into your IR, you have the added aggravation of having to
> re-write the code twice.  Once to convert it into SSA form and
> the second time to convert it out of SSA form.

Here's one possible way to look at this issue:  SIMPLE should always be in
SSA form.  When it is SIMPLE is initially constructed, SSA form is built.
When SIMPLE is lowered to RTL, UnSSA is run, or RTL in SSA form could even
be created and the existing RTL UnSSA pass could be used eventually.

> The approach we use now does not interfere with the underlying
> IR.  What it does is overlay a web of def-use/use-def pointers on

Is there a reason that you'd like to have a SIMPLE form that is not in
SSA?

-Chris

http://llvm.cs.uiuc.edu/
http://www.nondot.org/~sabre/Projects/