Re: [aarch64] Code-gen for vector initialization involving constants

[Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>]

On Fri, 3 Feb 2023 at 12:46, Prathamesh Kulkarni
<prathamesh.kulkarni@linaro.org> wrote:
>
> Hi Richard,
> While digging thru aarch64_expand_vector_init, I noticed it gives
> priority to loading a constant first:
>  /* Initialise a vector which is part-variable.  We want to first try
>      to build those lanes which are constant in the most efficient way we
>      can.  */
>
> which results in suboptimal code-gen for following case:
> int16x8_t f_s16(int16_t x)
> {
>   return (int16x8_t) { x, x, x, x, x, x, x, 1 };
> }
>
> code-gen trunk:
> f_s16:
>         movi    v0.8h, 0x1
>         ins     v0.h[0], w0
>         ins     v0.h[1], w0
>         ins     v0.h[2], w0
>         ins     v0.h[3], w0
>         ins     v0.h[4], w0
>         ins     v0.h[5], w0
>         ins     v0.h[6], w0
>         ret
>
> The attached patch tweaks the following condition:
> if (n_var == n_elts && n_elts <= 16)
>   {
>     ...
>   }
>
> to pass if maxv >= 80% of n_elts, with 80% being an
> arbitrary "high enough" threshold. The intent is to dup
> the most repeating variable if it it's repetition
> is "high enough" and insert constants which should be "better" than
> loading constant first and inserting variables like in the above case.
>
> Alternatively, I suppose we can remove threshold and for constants,
> generate both sequences and check which one is more
> efficient ?
>
> code-gen with patch:
> f_s16:
>         dup     v0.8h, w0
>         movi    v1.4h, 0x1
>         ins     v0.h[7], v1.h[0]
>         ret
>
> The patch is lightly tested to verify that vec[t]-init-*.c tests pass
> with bootstrap+test
> in progress.
> Does this look OK ?
Hi Richard,
ping https://gcc.gnu.org/pipermail/gcc-patches/2023-February/611243.html

Thanks,
Prathamesh
>
> Thanks,
> Prathamesh