This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
RE: i386 backend
- To: "Neil Booth" <neil at daikokuya dot demon dot co dot uk>
- Subject: RE: i386 backend
- From: "David Rasmussen" <pinkfloydhomer at yahoo dot com>
- Date: Tue, 20 Mar 2001 08:51:20 +0100
- Cc: <gcc at gcc dot gnu dot org>
> David Rasmussen wrote:-
>
> > I can understand that, but as I've explained elsewhere, the process of
> > isolating code for testcase purposes, is not necesarily a simple one,
> > especially not if the program in question has a complex
> interrelationship
> > between a number of critical functions.
> >
> > I was hoping that my observations, which are not concrete on
> purpose, could
> > stand for themselves, because people who _have_ isolated such
> code, and who
> > knows more about gcc that I do, would say: He has a point
> there. I've been
> > experiencing the same thing.
> >
> > But I guess no one here has, then.
>
> We know of weaknesses and strengths of the compiler, and we try to improve
> the weaknesses. If you can't come up with a concrete example, then this
> thread hasn't helped one bit. Just like me saying your chess program is
> not as good as Fritz in endgames, but without ever providing an example
> endgame.
>
It certain isn't as good as Fritz in any part of the game :)
If you want an instance where the difference is pronounced, compile Crafty,
which I have posted a link to before. You will see that Crafty is
significantly slower on gcc. If that isn't a testcase, I don't know what is.
I can understand the need for isolated small test examples, but it is not
necesarily possible to get that. That doesn't mean that the problem
shouldn't be recognized.
> Anyway, this has gone on enough.
>
What do you mean?
David.