|Are SSA and reaching definitions equivalent? email@example.com (Amal Banerjee) (2002-04-06)|
|Re: Are SSA and reaching definitions equivalent? firstname.lastname@example.org (2002-04-07)|
|From:||email@example.com (Andrew Bromage)|
|Date:||7 Apr 2002 00:48:30 -0500|
|Posted-Date:||07 Apr 2002 00:48:30 EST|
Amal Banerjee <firstname.lastname@example.org> writes:
> Are SSA and reaching definitions equivalent? If so what would be a
> formal way to prove/disprove it?
SSA form gives you more information than reaching definitions, namely,
the paths along which the definitions reach. That can be useful (e.g.
for constructing an interference graph, or for conditional constant
propagation), but otherwise I believe they're identical.
I haven't seen a formal proof, but I would think that proving that
reaching definition information can be obtained from SSA form should
be straightforward: you just trace the chains backwards through phi
nodes, after all.
Return to the
Search the comp.compilers archives again.