On Fri, 28 Jan 2022, Richard Biener wrote:
> > that's not what it does. It treats it like
> >
> > float tem = f;
> > return x + { tem, tem, tem, tem };
> >
> > avoiding, like for x + (1.0f + f) creating
> >
> > return x + { 1.0+f, 1.0+f, 1.0+f ...}
> >
> > it's more CSE than volatile qualif
On Fri, Jan 28, 2022 at 1:22 PM Richard Biener
wrote:
>
> On Thu, Jan 27, 2022 at 8:14 PM Maciej W. Rozycki wrote:
> >
> > On Thu, 27 Jan 2022, Richard Biener wrote:
> >
> > > > > > Index: gcc/gcc/c/c-typeck.cc
> > > > > > ===
> > >
On Thu, Jan 27, 2022 at 8:14 PM Maciej W. Rozycki wrote:
>
> On Thu, 27 Jan 2022, Richard Biener wrote:
>
> > > > > Index: gcc/gcc/c/c-typeck.cc
> > > > > ===
> > > > > --- gcc.orig/gcc/c/c-typeck.cc
> > > > > +++ gcc/gcc/c/c-typeck.c
On Thu, 27 Jan 2022, Richard Biener wrote:
> > > > Index: gcc/gcc/c/c-typeck.cc
> > > > ===
> > > > --- gcc.orig/gcc/c/c-typeck.cc
> > > > +++ gcc/gcc/c/c-typeck.cc
> > > > @@ -49,6 +49,7 @@ along with GCC; see the file COPYING3.
> >
On Thu, Jan 27, 2022 at 2:59 PM Maciej W. Rozycki wrote:
>
> On Thu, 27 Jan 2022, Richard Biener wrote:
>
> > > Index: gcc/gcc/c/c-typeck.cc
> > > ===
> > > --- gcc.orig/gcc/c/c-typeck.cc
> > > +++ gcc/gcc/c/c-typeck.cc
> > > @@ -49,6
On Thu, 27 Jan 2022, Richard Biener wrote:
> > Index: gcc/gcc/c/c-typeck.cc
> > ===
> > --- gcc.orig/gcc/c/c-typeck.cc
> > +++ gcc/gcc/c/c-typeck.cc
> > @@ -49,6 +49,7 @@ along with GCC; see the file COPYING3.
> > #include "gomp-cons
On Thu, Jan 27, 2022 at 1:16 PM Maciej W. Rozycki wrote:
>
> Do not mark vector element broadcast resulting from OpenCL operations as
> having side effects for targets that have a suitable hardware operation,
> so that the `vec_duplicateM' standard RTL pattern can be directly used
> for them. Thi
Do not mark vector element broadcast resulting from OpenCL operations as
having side effects for targets that have a suitable hardware operation,
so that the `vec_duplicateM' standard RTL pattern can be directly used
for them. This does not happen currently, because any RTL pattern named
`vec_
