This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: proposal for compilation unil wide alias analyis

Diego Novillo wrote:

On Fri, 2004-06-25 at 18:08, Jan Hubicka wrote:

However the mainline might be good as well as we already do have
functions gimplified at the time cgraph_analize is called (ie when you
would like to do the analysis).

Really?  Then why do we call the gimplifier as one of the tree
optimization passes?  That makes things easier then.  But the bodies are
still in high GIMPLE, are they not?  Meaning that they still have
nesting and scopes.

What would be Really Cool, though, is to have all the function bodies in
the call graph already in SSA form.

Now, when exactly is cgraph_analyze called?  After we called t_r_o_c on
each function body?  That'd be too late.  We want to do this analysis
before we optimize the first body in the call graph.

I will need someone's help to do this. I was not going to try to restructure the compiler on my first programming assignment, only insert small changes here and there to get the info I needed.

Also I was under impression that doing basic IPA alias analysis should
be just question of hooking Daniel's points to analyzers into
cgraph code

This would fix some of the low hanging fruit wrt IPA analysis.  I don't
think the idea is to do a full context-sensitive analysis at first.


This is not the goal of this first hack. My initial hack is remove the needlessly bad assumptions we are making about file level statics. For this I need to look the entire file at one time and gather some simple information. Diego felt that this should be done at the gimplified level because it is quite likely that we will eventually want to do something more aggressive and if we start to high, we will have to start from scratch.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]