[Bug tree-optimization/111131] New: SLP of gathers incomplete

[Bug tree-optimization/111131] New: SLP of gathers incomplete

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111131

            Bug ID: 111131
           Summary: SLP of gathers incomplete
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: rguenth at gcc dot gnu.org
  Target Milestone: ---

SLP of gathers relies on vect_recog_gather_scatter_pattern but for example
x86 still uses the decl based target hook.  There's also open-coded
gather/scatter support which isn't handled by SLP either.

The "correct" way is probably to use vect_check_gather_scatter at SLP
build time.

This can for example be seen when trying to vectorize
gcc.dg/vect/tsvc/vect-tsvc-vag.c with single-lane SLP, but
gcc.dg/vect/vect-gather-*.c also FAIL without vect_gather_load_ifn.

[Bug tree-optimization/111131] SLP of gathers incomplete

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111131

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
There's also

  if (memory_access_type == VMAT_GATHER_SCATTER && gs_info.decl)
    {
      vect_build_gather_load_calls (vinfo, stmt_info, gsi, vec_stmt, &gs_info,
                                    mask, cost_vec);
      return true;

which doesn't support SLP either (but ICEs).

[Bug tree-optimization/111131] SLP of gathers incomplete

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111131

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2023-10-19

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Have a nearly finished patch.

[Bug tree-optimization/111131] SLP of gathers incomplete

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111131

--- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Richard Biener <rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:beab5b95c581452adeb26efd59ae84a61fb3b429

commit r14-4745-gbeab5b95c581452adeb26efd59ae84a61fb3b429
Author: Richard Biener <rguenther@suse.de>
Date:   Thu Oct 19 10:33:01 2023 +0200

    tree-optimization/111131 - SLP for non-IFN gathers

    The following implements SLP vectorization support for gathers
    without relying on IFNs being pattern detected (and supported by
    the target).  That includes support for emulated gathers but also
    the legacy x86 builtin path.

            PR tree-optimization/111131
            * tree-vect-loop.cc (update_epilogue_loop_vinfo): Make
            sure to update all gather/scatter stmt DRs, not only those
            that eventually got VMAT_GATHER_SCATTER set.
            * tree-vect-slp.cc (_slp_oprnd_info::first_gs_info): Add.
            (vect_get_and_check_slp_defs): Handle gathers/scatters,
            adding the offset as SLP operand and comparing base and scale.
            (vect_build_slp_tree_1): Handle gathers.
            (vect_build_slp_tree_2): Likewise.

            * gcc.dg/vect/vect-gather-1.c: Now expected to vectorize
            everywhere.
            * gcc.dg/vect/vect-gather-2.c: Expected to not SLP anywhere.
            Massage the scale case to more reliably produce a different
            one.  Scan for the specific messages.
            * gcc.dg/vect/vect-gather-3.c: Masked gather is also supported
            for AVX2, but not emulated.
            * gcc.dg/vect/vect-gather-4.c: Expected to not SLP anywhere.
            Massage to more properly ensure this.
            * gcc.dg/vect/tsvc/vect-tsvc-s353.c: Expect to vectorize
            everywhere.

[Bug tree-optimization/111131] SLP of gathers incomplete

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111131

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
   Target Milestone|---                         |14.0
         Resolution|---                         |FIXED

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed.

[Bug tree-optimization/111131] SLP of gathers incomplete

[carlos.seo at linaro dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111131

Carlos Eduardo Seo <carlos.seo at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |carlos.seo at linaro dot org

--- Comment #5 from Carlos Eduardo Seo <carlos.seo at linaro dot org> ---
This patch caused a few extra failures:

                === gcc tests ===

Running gcc:gcc.target/aarch64/sve/aarch64-sve.exp ...
FAIL: gcc.target/aarch64/sve/mask_gather_load_8.c (internal compiler error: in
vect_get_and_check_slp_defs, at tree-vect-slp.cc:667)
FAIL: gcc.target/aarch64/sve/mask_gather_load_8.c (test for excess errors)
UNRESOLVED: gcc.target/aarch64/sve/mask_gather_load_8.c scan-assembler-times
\\tld1d\\tz[0-9]+\\.d, p[0-7]/z, \\[x[0-9]+, z[0-9]+\\.d, lsl #?3\\] 1
UNRESOLVED: gcc.target/aarch64/sve/mask_gather_load_8.c scan-assembler-times
\\tld1w\\tz[0-9]+\\.d, p[0-7]/z, \\[x[0-9]+, z[0-9]+\\.d, lsl #?2\\] 1
UNRESOLVED: gcc.target/aarch64/sve/mask_gather_load_8.c scan-assembler-times
\\tld1w\\tz[0-9]+\\.s, p[0-7]/z, \\[x[0-9]+, z[0-9]+\\.s, sxtw #?2\\] 1
UNRESOLVED: gcc.target/aarch64/sve/mask_gather_load_8.c scan-assembler-times
\\tld1w\\tz[0-9]+\\.s, p[0-7]/z, \\[x[0-9]+, z[0-9]+\\.s, uxtw #?2\\] 1

Could you please check?

[Bug tree-optimization/111131] SLP of gathers incomplete

[rguenther at suse dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111131

--- Comment #6 from rguenther at suse dot de <rguenther at suse dot de> ---
On Fri, 20 Oct 2023, carlos.seo at linaro dot org wrote:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111131
> 
> Carlos Eduardo Seo <carlos.seo at linaro dot org> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|                            |carlos.seo at linaro dot org
> 
> --- Comment #5 from Carlos Eduardo Seo <carlos.seo at linaro dot org> ---
> This patch caused a few extra failures:
> 
>                 === gcc tests ===
> 
> Running gcc:gcc.target/aarch64/sve/aarch64-sve.exp ...
> FAIL: gcc.target/aarch64/sve/mask_gather_load_8.c (internal compiler error: in
> vect_get_and_check_slp_defs, at tree-vect-slp.cc:667)
> FAIL: gcc.target/aarch64/sve/mask_gather_load_8.c (test for excess errors)
> UNRESOLVED: gcc.target/aarch64/sve/mask_gather_load_8.c scan-assembler-times
> \\tld1d\\tz[0-9]+\\.d, p[0-7]/z, \\[x[0-9]+, z[0-9]+\\.d, lsl #?3\\] 1
> UNRESOLVED: gcc.target/aarch64/sve/mask_gather_load_8.c scan-assembler-times
> \\tld1w\\tz[0-9]+\\.d, p[0-7]/z, \\[x[0-9]+, z[0-9]+\\.d, lsl #?2\\] 1
> UNRESOLVED: gcc.target/aarch64/sve/mask_gather_load_8.c scan-assembler-times
> \\tld1w\\tz[0-9]+\\.s, p[0-7]/z, \\[x[0-9]+, z[0-9]+\\.s, sxtw #?2\\] 1
> UNRESOLVED: gcc.target/aarch64/sve/mask_gather_load_8.c scan-assembler-times
> \\tld1w\\tz[0-9]+\\.s, p[0-7]/z, \\[x[0-9]+, z[0-9]+\\.s, uxtw #?2\\] 1
> 
> Could you please check?

That was reported already and fixed.