A thought about cc0 migration. (Re: How aggressive is CSE about

[Ulrich Weigand]

Kazu Hirata wrote:

> According to him, the first two are used mainly (only?) by the
> combiner.

Yes, they will be introduced by combine where appropriate.

> I'm wondering if it's a good idea to improve the combiner
> as follows.  When the "compare only" version fails to be recognized,
> fall back to the "compare and set" version.  This way, we can reduce
> the number of arithmetic insns by 1/3 or so.  

I don't recall exactly (it's already 3 years since I did the
CCmode conversion :-/), but I think that combine should already
act that way; the reason for the third insn was rather this:

> Of course, there is a
> possibility that passes after the combiner don't like the "compare and
> test" version very much.

It is still the case that a 'single_set' type instruction 
participates in a lot more passes than a parallel of two
sets, I think.

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  weigand@informatik.uni-erlangen.de