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: [tree-ssa] PATCH to improve optimization at ssa rewrite time

On Tue, 2003-05-13 at 16:48, Daniel Berlin wrote:
> On Tuesday, May 13, 2003, at 04:33  PM, Andrew MacLeod wrote:
> > On Tue, 2003-05-13 at 15:50, Daniel Berlin wrote:

> >
> > and then it seems pretty easy to do exactly the same thing with the
> > caclulation of b_7
> >
> > its clear that on edge (1,3)  b_7 can be calculated as 6 * 5, and on
> > edge (2,3) its 7 * 5...
> >
> > So  isn't it just as trivial for someone to insert temps on those 
> > edges:
> Who is this someone?
> Are we going to do out-of-ssa const-prop too?
> Or do you think CCP should be doing it?

I think CCP ought to be able to handle that pretty easily. At least the
all constant side of it.

> Diego also suggested adding temps in this case, my response was, of 
> course:
> Isn't it easier to make it not miss these optimizations in the first 
> place by not const-prop'ing when we can't const-prop it everywhere 
> (getting the trivial cases, but leaving the complex ones to passes that 
> handle it right), than it is to add passes and code in them to fix it 
> up?

right now, I dont know. I havent really had time to study it, and we
don't have enough working opimizations to do any real empirical studies
that are useful yet.

I definately tend to favour optimizations doing just their thing and
nothing else. They shouldn't be obfuscated by trying to do something
that just happens to be convienient. So I am a tad leery of these
into-ssa optimizations, but its pretty early too.  :-)

We'll just have to measure and experiment. So does PRE do anything when
these are turned on?


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