[Vectorizer] Add SLP support for masked loads

[Richard Biener]

On Wed, Jan 16, 2019 at 2:37 PM Alejandro Martinez Vicente
<Alejandro.MartinezVicente@arm.com> wrote:
>
> Hi,
>
> Current vectorizer doesn't support masked loads for SLP. We should add that, to
> allow things like:
>
> void
> f (int *restrict x, int *restrict y, int *restrict z, int n)
> {
>   for (int i = 0; i < n; i += 2)
>     {
>       x[i] = y[i] ? z[i] : 1;
>       x[i + 1] = y[i + 1] ? z[i + 1] : 2;
>     }
> }
>
> to be vectorized using contiguous loads rather than LD2 and ST2.
>
> This patch was motivated by SVE, but it is completely generic and should apply
> to any architecture with masked loads.
>
> After the patch is applied, the above code generates this output
> (-march=armv8.2-a+sve -O2 -ftree-vectorize):
>
> 0000000000000000 <f>:
>    0:   7100007f        cmp     w3, #0x0
>    4:   540002cd        b.le    5c <f+0x5c>
>    8:   51000464        sub     w4, w3, #0x1
>    c:   d2800003        mov     x3, #0x0                        // #0
>   10:   90000005        adrp    x5, 0 <f>
>   14:   25d8e3e0        ptrue   p0.d
>   18:   53017c84        lsr     w4, w4, #1
>   1c:   910000a5        add     x5, x5, #0x0
>   20:   11000484        add     w4, w4, #0x1
>   24:   85c0e0a1        ld1rd   {z1.d}, p0/z, [x5]
>   28:   2598e3e3        ptrue   p3.s
>   2c:   d37ff884        lsl     x4, x4, #1
>   30:   25a41fe2        whilelo p2.s, xzr, x4
>   34:   d503201f        nop
>   38:   a5434820        ld1w    {z0.s}, p2/z, [x1, x3, lsl #2]
>   3c:   25808c11        cmpne   p1.s, p3/z, z0.s, #0
>   40:   25808810        cmpne   p0.s, p2/z, z0.s, #0
>   44:   a5434040        ld1w    {z0.s}, p0/z, [x2, x3, lsl #2]
>   48:   05a1c400        sel     z0.s, p1, z0.s, z1.s
>   4c:   e5434800        st1w    {z0.s}, p2, [x0, x3, lsl #2]
>   50:   04b0e3e3        incw    x3
>   54:   25a41c62        whilelo p2.s, x3, x4
>   58:   54ffff01        b.ne    38 <f+0x38>  // b.any
>   5c:   d65f03c0        ret
>
>
> I tested this patch in an aarch64 machine bootstrapping the compiler and
> running the checks.

Thanks for implementing this - note this is stage1 material and I will have
a look when time allows unless Richard beats me to it.

It might be interesting to note that "non-SLP" code paths are likely to
go away in GCC 10 to streamline the vectorizer and make further changes
easier (so you'll see group_size == 1 SLP instances).

There are quite a few other cases missing SLP handling.

Richard.

> Alejandro
>
> gcc/Changelog:
>
> 2019-01-16  Alejandro Martinez  <alejandro.martinezvicente@arm.com>
>
>         * config/aarch64/aarch64-sve.md (copysign<mode>3): New define_expand.
>         (xorsign<mode>3): Likewise.
>         internal-fn.c: Marked mask_load_direct and mask_store_direct as
>         vectorizable.
>         tree-data-ref.c (data_ref_compare_tree): Fixed comment typo.
>         tree-vect-data-refs.c (can_group_stmts_p): Allow masked loads to be
>         combined even if masks different.
>         (slp_vect_only_p): New function to detect masked loads that are only
>         vectorizable using SLP.
>         (vect_analyze_data_ref_accesses): Mark SLP only vectorizable groups.
>         tree-vect-loop.c (vect_dissolve_slp_only_groups): New function to
>         dissolve SLP-only vectorizable groups when SLP has been discarded.
>         (vect_analyze_loop_2): Call vect_dissolve_slp_only_groups when needed.
>         tree-vect-slp.c (vect_get_and_check_slp_defs): Check masked loads
>         masks.
>         (vect_build_slp_tree_1): Fixed comment typo.
>         (vect_build_slp_tree_2): Include masks from masked loads in SLP tree.
>         tree-vect-stmts.c (vect_get_vec_defs_for_operand): New function to get
>         vec_defs for operand with optional SLP and vectype.
>         (vectorizable_load): Allow vectorizaion of masked loads for SLP only.
>         tree-vectorizer.h (_stmt_vec_info): Added flag for SLP-only
>         vectorizable.
>         tree-vectorizer.c (vec_info::new_stmt_vec_info): Likewise.
>
> gcc/testsuite/Changelog:
>
> 2019-01-16  Alejandro Martinez  <alejandro.martinezvicente@arm.com>
>
>         * gcc.target/aarch64/sve/mask_load_slp_1.c: New test for SLP
>         vectorized masked loads.