Re: SSA-form and cmove

hannah@mamba.pond.sub.org (Hannah Schroeter)
13 Sep 2000 21:10:39 -0400

          From comp.compilers

Related articles
SSA-form and cmove tommy.hoffner@softlab.ericsson.se (Tommy Hoffner) (2000-08-13)
Re: SSA-form and cmove stephen.clarke@earthling.net (Stephen Clarke) (2000-08-20)
Re: SSA-form and cmove tlh20@cam.ac.uk (Tim Harris) (2000-08-27)
Re: SSA-form and cmove krish@wam.umd.edu (2000-09-02)
Re: SSA-form and cmove hannah@mamba.pond.sub.org (2000-09-13)
| List of all articles for this month |
From: hannah@mamba.pond.sub.org (Hannah Schroeter)
Newsgroups: comp.compilers
Date: 13 Sep 2000 21:10:39 -0400
Organization: Chaos
References: 00-08-066
Keywords: analysis

Hello!


Tommy Hoffner <tommy.hoffner@softlab.ericsson.se> wrote:
>[... conditional move vs. SSA ...]


Couldn't this be handled like just another operator, just like
plus or something like that?


i.e. (like v_new := v_old1 + v_old2)
v_new:= condition ? v_old1 : v_old2


(and cmove dest, src, condition is just
dest_new:= condition ? src : dest_old with a desire to
coalesce dest_new and dest_old).


In this view, I don't see the need for a phi-node there, however,
transformations should know that if they can prove condition is
true, that's equivalent to dest_new := src (and dest_old is dead),
and if condition is false, that's dest_new := dest_old (and src is dead).


Kind regards,


Hannah.


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.