Re: Reload problem

[Geoff Keating <geoffk at geoffk dot org>]

Joern Rennecke <joern.rennecke@superh.com> writes:

> > On Thu, 2003-09-11 at 02:09, Eric Botcazou wrote:
> > > It is my understanding that recognizing insns is an expensive process, so I 
> > > think that adding it to push_reload would be the last resort solution.
> 
> Jim Wilson:
> > Yes, it is expensive, so we want to avoid doing it too often.  It is
> > however the only test that exactly describes the condition we are
> > looking for.  Anything else will require compromising in some cases.
> 
> I think the cost is not that significant, since we are only doing this
> check when we are likely to emit the instruction we are trying to
> recognize, which then has to be emitted and recognized, too.
> 
> If we are seriously concerned about performance there, we can cache the
> insn code or even the entire emitted instruction for use by gen_reload.

I think that would be a good idea, if only to reduce stress on the
memory manager.

-- 
- Geoffrey Keating <geoffk@geoffk.org>