[PATCH][x86] Add more patterns for SSE/AVX scalar single/double-precision fp arithmetic instructions.

Andrea Di Biagio andrea.dibiagio at gmail.com
Wed Dec 11 13:14:33 PST 2013

Thanks Nadav.

I will then commit this patch first and as a next step I'll start
working on how to canonicalize the shuffle into a blend.


On Wed, Dec 11, 2013 at 8:47 PM, Nadav Rotem <nrotem at apple.com> wrote:
> On Dec 11, 2013, at 12:44 PM, Andrea Di Biagio <andrea.dibiagio at gmail.com>
> wrote:
> With that combine, the new patterns would work even when a 'vselect' is
> used.
> What is your opinion about this?
> Sounds good.
> In case, if you think that this is the correct way to fix it, would it
> be acceptable to commit the current patch first and then continue
> working on the missing combine?
> Yes, we can start with the current patterns, but I would like us to
> canonicalize the shuffle into a blend as a next step.
> Thanks,
> Nadav

More information about the llvm-commits mailing list