This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: VIA C3-2 support.
- From: Jan Hubicka <jh at suse dot cz>
- To: Dave Jones <davej at codemonkey dot org dot uk>, gcc-patches at gcc dot gnu dot org
- Date: Sun, 9 Feb 2003 19:18:31 +0100
- Subject: Re: VIA C3-2 support.
- References: <20030209150728.GA12877@codemonkey.org.uk>
> The Nehemiah (model 9) core of the VIA C3 is significantly different
> from the earlier C3s. It now has cmov (so we can generate i686
> code for it now), and it dropped 3dnow, and implements SSE instead.
> This means that the existing -march=c3 will generate code that the
> C3-2 cannot run (and vice versa)
>
> Simple benchmarks on the C3-2 show that scheduling for 686 offers greater
> improvement than scheduling for 486 like on previous C3s.
Hmm, perhaps we can cook up some c3 specific scheudling?
I don't like much the idea of simply aliasing the CPU to something else
much. Adding new type for it would let you to define costs and
instruction selection (see tables in the front of i386.c)
>
> I'm uncertain my choice to alias target_cpu_default2=TARGET_CPU_DEFAULT_pentium3
> is correct, but it seems the closest match, with minimal code addition.
>
> I've bootstrapped and tested this on i686-linux, and everything
> seems to be as expected.
>
> Patch below is based upon Jeff Garziks original that got merged.
> As before with Jeff, I don't have a copyright assignment on file,
> but I'm prepared to offer this patch into public domain and let
> anyone do what they like with it.
You can't, as your code is infected by GPL. You really need assignment
for anything nontrivial (perhaps this patch is trivial)
Honza