This is the mail archive of the
mailing list for the GCC project.
Re: Patch to fix register allocation bug
- From: Richard Sandiford <rsandifo at redhat dot com>
- To: Michael Matz <matz at suse dot de>
- Cc: <gcc-patches at gcc dot gnu dot org>
- Date: 18 Nov 2002 14:30:53 +0000
- Subject: Re: Patch to fix register allocation bug
- References: <Pine.LNX.email@example.com>
Michael Matz <firstname.lastname@example.org> writes:
> I now think for a hot fix it's good enough (but I've no karma). The right
> one would be to teach it about multi-reg pseudos, but well, that would
> need a loop to cumulate the costs of all parts and it's not clear this is
> worth anything ;-)
I wondered about that, but I wasn't sure whether simply adding
the costs of each individual register would be good enough.
Would it accurately trade off one DFmode load or store against
two SFmode ones?
It's probably worth noting that, as well as fixing the
original test case, the patch gave significantly smaller
output. It was the kind of code that had a lot of printf()s
so I'm afraid I didn't bother measuring the speed difference.