|renaming question firstname.lastname@example.org (Xinan TANG) (1994-09-29)|
|Re: renaming question email@example.com (1994-10-02)|
|From:||Xinan TANG <firstname.lastname@example.org>|
|Keywords:||optimize, question, comment|
|Date:||Thu, 29 Sep 1994 21:55:26 GMT|
I am just wondering whether someone has already done something which I
am thinking to start to do.
The question is that given a C program, is it possible to do some
renaming to either expose parallelism or facilitate register allocation to
At the first step, it might be easy if only SCALAR variable is
considered. Is there any efficient algorithm to do this sort of thing?
When talking about renaming, I mean to do source to source
transformation and want to keep the original semantics of the program.
Please don't point me some thing like SSA beacuse that's not I want.
[Loop unrolling is certainly this kind of transformation. -John]
Return to the
Search the comp.compilers archives again.