Re: [00/13] Make VEC_PERM_EXPR work for variable-length vectors

[Richard Biener <richard dot guenther at gmail dot com>]

On Sun, Dec 10, 2017 at 12:06 AM, Richard Sandiford
<richard.sandiford@linaro.org> wrote:
> This series is a replacement for:
> https://gcc.gnu.org/ml/gcc-patches/2017-11/msg00747.html
> based on the feedback that using VEC_PERM_EXPR would be better.
>
> The changes are:
>
> (1) Remove the restriction that the selector elements have to have the
>     same width as the data elements, but only for constant selectors.
>     This lets through the cases we need without also allowing
>     potentially-expensive ops.  Adding support for the variable
>     case can be done later if it seems useful, but it's not trivial.
>
> (2) Encode the integer form of constant selectors (vec_perm_indices)
>     in the same way as the new VECTOR_CST encoding, so that it can
>     cope with variable-length vectors.
>
> (3) Remove the vec_perm_const optab and reuse the target hook to emit
>     code.  This avoids the need to create a CONST_VECTOR for the wide
>     selectors, and hence the need to have a corresponding wide vector
>     mode (which the target wouldn't otherwise need or support).

Hmm.  Makes sense I suppose.

> (4) When handling the variable vec_perm optab, check that modes can store
>     all element indices before using them.
>
> (5) Unconditionally use ssizetype selector elements in permutes created
>     by the vectoriser.

Why specifically _signed_ sizetype?  That sounds like an odd choice.  But I'll
eventually see when looking at the patch.  Does that mean we have a
VNDImode vector unconditionally for the permute even though a vector
matching the width of the data members would work?  What happens if the
target doesn't have vec_perm_const but vec_perm to handle all constant
permutes?

Going to look over the patches now.

Thanks for (re-)doing the work.

Richard.

> (6) Make the AArch64 vec_perm_const handling handle variable-length vectors.
>
> Tested directly on trunk on aarch64-linux-gnu, x86_64-linux-gnu and
> powerpc64le-linux-gnu.  Also tested by comparing the before and after
> assembly output for:
>
>    arm-linux-gnueabi arm-linux-gnueabihf aarch64-linux-gnu
>    aarch64_be-linux-gnu ia64-linux-gnu i686-pc-linux-gnu
>    mipsisa64-linux-gnu mipsel-linux-gnu powerpc64-linux-gnu
>    powerpc64le-linux-gnu powerpc-eabispe x86_64-linux-gnu
>    sparc64-linux-gnu
>
> at -O3, which should cover all the ports that defined vec_perm_const.
> The only difference was one instance of different RA for ia64-linux-gnu,
> caused by using force_reg on a SUBREG that was previously used directly.
>
> OK to install?
>
> Thanks,
> Richard