[Bug tree-optimization/115383] New: ICE with TCVC_2 build

[Bug tree-optimization/115383] New: ICE with TCVC_2 build

[kugan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

            Bug ID: 115383
           Summary: ICE with TCVC_2 build
           Product: gcc
           Version: 15.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: kugan at gcc dot gnu.org
  Target Milestone: ---

Patch [PATCH 1/4] Relax COND_EXPR reduction vectorization SLP restriction  seem
to cause ICE while building TSVC_2

Reduced test:

cat tsvc_vec.i
void dummy();
void s331() {
  int j;
  for (int i; i; i++)
    if ((float)i < 00.)
      j = i;
  dummy(j);
}


gcc options used:
gcc  -std=c99 -O3 -march=native -flto -Wl,--sort-section=name -mcpu=neoverse-v2
-msve-vector-bits=128
gcc -v:
Using built-in specs.
COLLECT_GCC=/proj/grco/gcc/Linux_aarch64/upstream-main/latest/bin/gcc
COLLECT_LTO_WRAPPER=/proj/grco/gcc/Linux_aarch64/upstream-main/20240606024711346f33e2/libexec/gcc/aarch64-unknown-linux-gnu/15.0.0/lto-wrapper
Target: aarch64-unknown-linux-gnu
Configured with: /var/jenkins/workspace/GCC_Nightly/configure
--enable-multiarch=yes --enable-languages=c,c++,fortran,lto
--prefix=/proj/grco/gcc/Linux_aarch64/upstream-main/20240606024711346f33e2
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 15.0.0 20240606 (experimental) (GCC)

[Bug tree-optimization/115383] [15 Regression] ICE with TCVC_2 build

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2024-06-07
                 CC|                            |ktkachov at gcc dot gnu.org
             Target|                            |aarch64
   Target Milestone|---                         |15.0
            Summary|[15 regression] ICE with    |[15 Regression] ICE with
                   |TCVC_2 build since          |TCVC_2 build
                   |r15-1053-g28edeb1409a7b8    |
      Known to fail|                            |15.0
     Ever confirmed|0                           |1
      Known to work|                            |14.1.0
           Keywords|                            |ice-on-valid-code

--- Comment #1 from ktkachov at gcc dot gnu.org ---
Confirmed. -O3 -msve-vector-bits=128 -mcpu=neoverse-v2 is enough to trigger it:
ice.c:2:6: internal compiler error: Segmentation fault
    2 | void s331() {
      |      ^~~~
0xf4b59b crash_signal
        $TOP/gcc/gcc/toplev.cc:319
0xb7f518 phi_nodes_ptr(basic_block_def*)
        $TOP/gcc/gcc/gimple.h:4701
0xb7f518 gsi_start_phis(basic_block_def*)
        $TOP/gcc/gcc/gimple-iterator.cc:937
0xb7f518 gsi_for_stmt(gimple*)
        $TOP/gcc/gcc/gimple-iterator.cc:621
0x1fa982f vectorizable_condition
        $TOP/gcc/gcc/tree-vect-stmts.cc:12577
0x1fc4627 vect_transform_stmt(vec_info*, _stmt_vec_info*,
gimple_stmt_iterator*, _slp_tree*, _slp_instance*)
        $TOP/gcc/gcc/tree-vect-stmts.cc:13467
0x1261733 vect_schedule_slp_node
        $TOP/gcc/gcc/tree-vect-slp.cc:9729
0x1276837 vect_schedule_slp_node
        $TOP/gcc/gcc/tree-vect-slp.cc:9522
0x1276837 vect_schedule_scc
        $TOP/gcc/gcc/tree-vect-slp.cc:10017
0x12776df vect_schedule_slp(vec_info*, vec<_slp_instance*, va_heap, vl_ptr>
const&)
        $TOP/gcc/gcc/tree-vect-slp.cc:10110
0x1244837 vect_transform_loop(_loop_vec_info*, gimple*)
        $TOP/gcc/gcc/tree-vect-loop.cc:12114
0x1287d5f vect_transform_loops
        $TOP/gcc/gcc/tree-vectorizer.cc:1007
0x12883e7 try_vectorize_loop_1
        $TOP/gcc/gcc/tree-vectorizer.cc:1153
0x12883e7 try_vectorize_loop
        $TOP/gcc/gcc/tree-vectorizer.cc:1183
0x128875b execute
        $TOP/gcc/gcc/tree-vectorizer.cc:1299
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

[Bug tree-optimization/115383] [15 Regression] ICE with TCVC_2 build

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org
           Priority|P3                          |P1

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

[Bug tree-optimization/115383] [15 Regression] ICE with TCVC_2 build since r15-1053-g28edeb1409a7b8

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rsandifo at gcc dot gnu.org

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
So we're now doing a EXTRACT_LAST_REDUCTION with multiple stmt copies which
is disallowed for non-SLP (by accident?).  It shows it of course doesn't
work since we end up removing the scalar reduction stmt multiple times.

  <bb 3> [local count: 860067202]:
  # j_12 = PHI <j_2(8), j_5(D)(28)>
  # i_14 = PHI <i_9(8), i_6(D)(28)>
  # vect_vec_iv_.9_45 = PHI <_46(8), _47(28)>
  _46 = vect_vec_iv_.9_45 + { 16, 16, 16, 16 };
  _48 = vect_vec_iv_.9_45 + { 4, 4, 4, 4 };
  _49 = _48 + { 4, 4, 4, 4 };
  _50 = _49 + { 4, 4, 4, 4 };
  vect__1.10_51 = (vector(4) float) vect_vec_iv_.9_45;
  vect__1.10_52 = (vector(4) float) _48;
  vect__1.10_53 = (vector(4) float) _49;
  vect__1.10_54 = (vector(4) float) _50;
  mask__3.11_55 = vect__1.10_51 < { 0.0, 0.0, 0.0, 0.0 };
  mask__3.11_56 = vect__1.10_52 < { 0.0, 0.0, 0.0, 0.0 };
  mask__3.11_57 = vect__1.10_53 < { 0.0, 0.0, 0.0, 0.0 };
  mask__3.11_58 = vect__1.10_54 < { 0.0, 0.0, 0.0, 0.0 };
  j_2 = .FOLD_EXTRACT_LAST (j_12, mask__3.11_55, vect_vec_iv_.9_45);

and we removed the old

  j_2 = _3 ? i_14 : j_12;

we are about to insert

  j_2 = .FOLD_EXTRACT_LAST (j_12, mask__3.11_56, _48);

I think correct would be

  j_59 = .FOLD_EXTRACT_LAST (j_12, mask__3.11_55, vect_vec_iv_.9_45);
  j_60 = .FOLD_EXTRACT_LAST (j_59, mask__3.11_56, _48);
  j_61 = .FOLD_EXTRACT_LAST (j_60, mask__3.11_57, _49);
  j_2 = .FOLD_EXTRACT_LAST (j_61, mask__3.11_58, _50);

I'm testing a patch.

[Bug tree-optimization/115383] [15 Regression] ICE with TCVC_2 build since r15-1053-g28edeb1409a7b8

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
Created attachment 58378
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=58378&action=edit
patch

I'm testing this, but I do not have hardware to test correctness (and qemu not
set up).

[Bug tree-optimization/115383] [15 Regression] ICE with TCVC_2 build since r15-1053-g28edeb1409a7b8

[kugan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

--- Comment #5 from kugan at gcc dot gnu.org ---
(In reply to Richard Biener from comment #4)
> Created attachment 58378 [details]
> patch
> 
> I'm testing this, but I do not have hardware to test correctness (and qemu
> not set up).

Thanks. I will test this on aarch64.

[Bug tree-optimization/115383] [15 Regression] ICE with TCVC_2 build since r15-1053-g28edeb1409a7b8

[kugan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

--- Comment #6 from kugan at gcc dot gnu.org ---
(In reply to kugan from comment #5)
> (In reply to Richard Biener from comment #4)
> > Created attachment 58378 [details]
> > patch
> > 
> > I'm testing this, but I do not have hardware to test correctness (and qemu
> > not set up).
> 
> Thanks. I will test this on aarch64.

bootstrap and regression test passes. TSVC_2 also builds without any issues.

[Bug tree-optimization/115383] [15 Regression] ICE with TCVC_2 build since r15-1053-g28edeb1409a7b8

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

--- Comment #7 from GCC 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:c1429e3a8da0cdfe9391e1e9b2c7228d896a3a87

commit r15-1126-gc1429e3a8da0cdfe9391e1e9b2c7228d896a3a87
Author: Richard Biener <rguenther@suse.de>
Date:   Fri Jun 7 12:15:31 2024 +0200

    tree-optimization/115383 - EXTRACT_LAST_REDUCTION with multiple stmt copies

    The EXTRACT_LAST_REDUCTION code isn't ready to deal with multiple stmt
    copies but SLP no longer checks for this.  The following adjusts
    code generation to handle the situation.

            PR tree-optimization/115383
            * tree-vect-stmts.cc (vectorizable_condition): Handle
            generating a chain of .FOLD_EXTRACT_LAST.

            * gcc.dg/vect/pr115383.c: New testcase.

[Bug tree-optimization/115383] [15 Regression] ICE with TCVC_2 build since r15-1053-g28edeb1409a7b8

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115383

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

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

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
Should be fixed now.