[PATCH] Improve XMM16-31 handling in various *vec_dup* patterns
Kirill Yukhin
kirill.yukhin@gmail.com
Sun May 22 07:39:00 GMT 2016
Hello Jakub,
On 18 May 22:58, Jakub Jelinek wrote:
> Hi!
>
> These instructions are available in AVX512VL, so we can use
> XMM16+ in there.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
OK.
>
> 2016-05-18 Jakub Jelinek <jakub@redhat.com>
>
> * config/i386/sse.md (avx2_vec_dupv4df): Use v instead of x
> constraint, use maybe_evex prefix instead of vex.
> (vec_dupv4sf): Use v constraint instead of x for output
> operand except for noavx alternative, use Yv constraint
> instead of x for input. Use maybe_evex prefix instead of vex.
> (*vec_dupv4si): Likewise.
> (*vec_dupv2di): Likewise.
>
> * gcc.target/i386/avx512vl-vbroadcast-1.c: New test.
--
Thanks, K
More information about the Gcc-patches
mailing list