Re: sub-optimal code for packed boolean arrays -- bug or inherent limitation

[Georg Bauhaus <rm.tsoh+bauhaus@maps.futureapps.de>]

On Tue, 2007-07-03 at 00:22 -0700, Harald Korneliussen wrote:
> On Jul 3, 3:01 am, "Jeffrey R. Carter"
> <spam.jrcarter....@acm.nospam.org> wrote:
> > I doubt the compiler's code will make your chess
> > program fail to meet its timing requirements; I doubt if you'll notice
> > the difference at all, so I don't see why you care. If you must have a
> > specific sequence of machine code operations, you should use a
> > machine-code insertion.
> >
> Now now, chess programs written to compete with other chess programs
> need all the performance they can get. You could argue that he should
> be using C with inline assembly (and probably another compiler than
> GCC), but wouldn't it be kind of neat if a competitive and readable
> chess program was written in Ada?

Clear Ada is the point of this Ada program as the OP has explained
on the GCC mailing list.


> I would ask the GCC team, after looking at the intermediate code. I'm
> sure they're not content with merely having the compiler generate
> correct code.

The suggestion was different. Still, the problem is identified
as non-optimal code.