On Fri, 18 Oct 2019, Christophe Lyon wrote:
> On Wed, 16 Oct 2019 at 15:09, Richard Biener wrote:
> >
> >
> > It happens we cannot have different typed data and index for
> > integer condition reductions right now, for whatever reason.
> > The following makes that work, even for double data and i
On Wed, 16 Oct 2019 at 15:09, Richard Biener wrote:
>
>
> It happens we cannot have different typed data and index for
> integer condition reductions right now, for whatever reason.
> The following makes that work, even for double data and integer index.
> There's hope this enables some relevant a
It happens we cannot have different typed data and index for
integer condition reductions right now, for whatever reason.
The following makes that work, even for double data and integer index.
There's hope this enables some relevant amount of extra vectorization.
Actually this is fallout from si
