Re: Combine four insns

[Eric Botcazou <ebotcazou at adacore dot com>]

> Thanks for doing this.  I think the patch is ok if you guard the
> 4-way combines by flag_expensive_optimizations.

Combining Steven and Bernd's figures, 1% of a bootstrap time is 37% of the 
combiner's time.  The result is 0.18% more combined insns.  It seems to me 
that we are already very far in the directory of diminishing returns.

Needless to say that, by fixing the problems earlier, we not only save 1% of 
compilation time but speed up the compiler by not generating the dreadful RTL 
in the first place and carrying it over to the combiner.

Bernd is essentially of the opinion that compilation time doesn't matter.  It 
seems to me that, even if we were to adopt this position, this shouldn't mean 
wasting compilation time, which I think is the case here.

So I think that the patch shouldn't go in at this point.

-- 
Eric Botcazou