On Tue, Nov 9, 2021 at 4:40 PM Bill Schmidt <wschm...@linux.ibm.com> wrote:
>
> Hi! Over the last month or so, Haochen made a couple of changes to the
> builtins
> support that need to be reflected into the new builtin support:
>
> 14e355df Disable gimple folding for vector min/max without fast-math
> 91419baf Optimize vec_xl_sext
>
> In both cases these require almost identical changes in the new builtins
> support
> to what was done for the old support. While testing, I also observed that
> Haochen
> made a small mistake in the test case gcc.target/powerpc/p10_vec_xl_sext.c,
> using
> "vector long" when "vector long long" is the officially supported interface.
> I
> corrected that as part of this patch.
>
> Bootstrapped and tested on powerpc64le-linux-gnu with no regressions. Is this
> okay for trunk?
>
> Thanks!
> Bill
>
>
> 2021-11-09 Bill Schmidt <wschm...@linux.ibm.com>
>
> gcc/
> * config/rs6000/rs6000-call.c (rs6000_gimple_fold_new_builtin):
> Disable gimple fold for RS6000_BIF_{XVMINDP,XVMINSP,VMINFP} and
> RS6000_BIF_{XVMAXDP,XVMAXSP,VMAXFP} when fast-math is not set.
> (lxvrse_expand_builtin): Modify the expansion for sign extension.
> All extensions are done within VSX registers.
>
> gcc/testsuite/
> * gcc.target/powerpc/p10_vec_xl_sext.c: Fix long long case.
Okay.
Thanks, David
