Re: open64 versus gcc

Gary Oblock <>
3 Dec 2006 21:34:35 -0500

          From comp.compilers

Related articles
[7 earlier articles]
Re: open64 versus gcc (A.L.) (2006-12-01)
Re: open64 versus gcc (A.L.) (2006-12-01)
Re: open64 versus gcc (Jonathan Thornburg -- remove -animal to reply) (2006-12-03)
Re: open64 versus gcc (Diego Novillo) (2006-12-03)
Re: open64 versus gcc (Greg Lindahl) (2006-12-03)
Re: open64 versus gcc (Brooks Moses) (2006-12-03)
Re: open64 versus gcc (Gary Oblock) (2006-12-03)
Re: open64 versus gcc (ST) (2006-12-06)
| List of all articles for this month |

From: Gary Oblock <>
Newsgroups: comp.compilers
Date: 3 Dec 2006 21:34:35 -0500
Organization: EarthLink Inc. --
References: 06-11-09406-11-100 06-11-104 06-11-113 06-11-120 06-11-124
Keywords: GCC, arithmetic
Posted-Date: 03 Dec 2006 21:34:35 EST

Diego Novillo wrote:

> dz wrote on 11/29/06 00:52:
>> The criteria I am looking for is stability and the strength of some
>> basic analysis in the compiler such as alias analysis. Can anyone
>> comment on that ?
> For alias analysis, GCC uses a fairly sophisticated constraint-based
> points-to analysis and complements it with type-based disambiguation.
> You can read about it in the various GCC Summit proceedings over the
> last 2-3 years. You can find them in
> I'm not sure what exactly you mean by "some basic analysis", but you
> will find GCC a fairly featureful compiler. In terms of strength, GCC
> is the system compiler of every Linux distribution out there, so it is
> thoroughly tested and breaking it takes a bit of effort.

I was worked on a custom VLIW scheduler grafted into gcc 4.x and at the
RTL level the alias information available wasn't all that great in my
opinion. I think there is a fundementail lack of communication between
the tree level and the RTL levels because though the initial RTL
generated has a way to get at the tree level information the subsequent
optimizations need to preserve it (which they don't.)

-- Gary

Bronze Dreams

Post a followup to this message

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