Re: Regarding transforming Control Flow Graph

"Aaron Gray" <angray@beeb.net>
16 Apr 2005 11:12:05 -0400

          From comp.compilers

Related articles
Regarding transforming Control Flow Graph vipin.v.deep@gmail.com (Deep.V...@gmail.com) (2005-04-11)
Re: Regarding transforming Control Flow Graph angray@beeb.net (Aaron Gray) (2005-04-16)
| List of all articles for this month |

From: "Aaron Gray" <angray@beeb.net>
Newsgroups: comp.compilers
Date: 16 Apr 2005 11:12:05 -0400
Organization: Customer of PlusNet plc (http://www.plus.net)
References: 05-04-036
Keywords: analysis
X-RFC2646: Format=Flowed; Original

> I am doing a project where I have to modify the program so that only
> subset of its path are retained and remove rest of the portions.
>
> For this I have to modify the control flow graph retaining necessary
> paths and pruning rest. What are the suffecient criterion to say that
> the transformations are legible. Some of them which I can think are-
> retaining dependencies, ensure return values.
>
> Is there any criterion which I have to prove (to be valid), for arguing
> that my transformations are correct.


This may help, have a look at the "Static Single Assignment" lecture
presentation :-


        http://www.cs.ualberta.ca/~amaral/courses/680/


Formal proof is another matter though.


Aaron


Post a followup to this message

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