Problems with the way we calculate costs

[Geoff Keating]

> Date: Fri, 8 Sep 2000 14:57:34 +0100 (BST)
> From: Bernd Schmidt <bernds@redhat.co.uk>

> I can think of several ways to fix that.  One of them is the patch below.
> I'm not too happy with it, but none of the other solutions I can think of
> are any more elegant.  Any better ideas?

It'd probably be better if it worked like the original code; that is,
the cost of a non-fixed hard register is 2, compared to 1 for a
pseudo.  This ensures a pseudo gets preference, which would be the
usual case because such hard registers should be copied into pseudos
immediately after they appear.

-- 
- Geoffrey Keating <geoffk@cygnus.com>