[Bug rtl-optimization/111142] New: [14 regression] ICE in get_group_load_store_type, at tree-vect-stmts.cc:2121 (AVX512)

[Bug rtl-optimization/111142] New: [14 regression] ICE in get_group_load_store_type, at tree-vect-stmts.cc:2121 (AVX512)

[manuel.lauss at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111142

            Bug ID: 111142
           Summary: [14 regression] ICE in get_group_load_store_type, at
                    tree-vect-stmts.cc:2121 (AVX512)
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: manuel.lauss at googlemail dot com
  Target Milestone: ---

Created attachment 55787
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=55787&action=edit
compressed unreduced testcase

gcc version 14.0.0 20230824 (experimental)
721f7e2c4e5eed645593258624dd91e6c39f3bd2


g++ -m32 -O2 -march=znver4 -c gcc-14-mesa-aco-ice.ii 
during GIMPLE pass: vect
../mesa-9999/src/amd/compiler/aco_insert_waitcnt.cpp: In function 'void
aco::insert_wait_states(Program*)':
../mesa-9999/src/amd/compiler/aco_insert_waitcnt.cpp:1042:1: internal compiler
error: in get_group_load_store_type, at tree-vect-stmts.cc:2121


can only trigger with "-march=znver4", does not trigger with avx2 only or -O1
or lower.

Fails with the same error for x64-64, but the testcase is preprocessed for
32bit targets.

Thank you!
      Manuel

[Bug rtl-optimization/111142] [14 regression] ICE in get_group_load_store_type, at tree-vect-stmts.cc:2121 (AVX512)

[manuel.lauss at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111142

--- Comment #1 from Manuel Lauss <manuel.lauss at googlemail dot com> ---
This is fairly new, something after 4aa14ec7d5b25722e4d02c29c8c1e22dcc5a4915
("MATCH: [PR111109] Fix bit_ior(cond,cond) when comparisons are fp") may have
caused this, since I was able to build mesa just 12 hours ago.

[Bug tree-optimization/111142] [14 regression] ICE in get_group_load_store_type, at tree-vect-stmts.cc:2121 (AVX512)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111142

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|rtl-optimization            |tree-optimization
           Keywords|                            |ice-on-valid-code
   Target Milestone|---                         |14.0

[Bug tree-optimization/111142] [14 regression] ICE in get_group_load_store_type, at tree-vect-stmts.cc:2121 (AVX512)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111142

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2023-08-25
             Status|UNCONFIRMED                 |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---

2120      /* Stores can't yet have gaps.  */
2121      gcc_assert (slp_node || vls_type == VLS_LOAD || gap == 0);
(gdb) p vls_type
$1 = VLS_STORE_INVARIANT

Looks like mine I think.  Reducing.

[Bug tree-optimization/111142] [14 regression] ICE in get_group_load_store_type, at tree-vect-stmts.cc:2121 (AVX512)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111142

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |DUPLICATE
             Status|ASSIGNED                    |RESOLVED

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
The fix for PR111136 also fixes this.

*** This bug has been marked as a duplicate of bug 111136 ***