[Bug tree-optimization/103317] New: Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[Bug tree-optimization/103317] New: Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[meissner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

            Bug ID: 103317
           Summary: Spec 2017 benchmark blender_r fails with -Ofast on
                    PowerPc (power9, power10)
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: meissner at gcc dot gnu.org
  Target Milestone: ---

Created attachment 51832
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51832&action=edit
Preprocessed .i file that shows the bug (gzipped).

As of November 17th, the current trunk compiler for GCC 12 fails when compiling
the Spec 2017 benchmark blender_r for PowerPC using -Ofast.  I see failures for
both power9 and power10 systems.  Here is the traceback:

libpng/png.c: In function 'png_build_gamma_table':
libpng/png.c:2872:1: error: definition in block 19 does not dominate use in
block 18
 2872 | }
      | ^
for SSA_NAME: _57 in statement:
shift_48 = PHI <_57(18), _58(19)>
PHI argument
_57
for PHI node
shift_48 = PHI <_57(18), _58(19)>
during GIMPLE pass: phiopt
libpng/png.c:2872:1: internal compiler error: verify_ssa failed
0x10c5f64b verify_ssa(bool, bool)
        /home/meissner/fsf-src/trunk-2021-11-17/gcc/tree-ssa.c:1211
0x1080e97f execute_function_todo
        /home/meissner/fsf-src/trunk-2021-11-17/gcc/passes.c:2049
0x1080f613 execute_todo
        /home/meissner/fsf-src/trunk-2021-11-17/gcc/passes.c:2096
Please submit a full bug report,
with preprocessed source if appropriate.

I see it with -O2, -O3, and -Ofast, using -mcpu=power8, -mcpu=power9, and
-mcpu=power10 on PowerPC little endian systems. It did compile with the
compiler I checked out on November 4th.

I will attach the .i file compressed with gzip.

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[meissner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

Michael Meissner <meissner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |major
           Priority|P3                          |P2

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #1 from Peter Bergner <bergner at gcc dot gnu.org> ---
I'll try and creduce the test case.

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[meissner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

Michael Meissner <meissner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P2                          |P1

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-11-18
           Keywords|                            |ice-on-valid-code

--- Comment #2 from Peter Bergner <bergner at gcc dot gnu.org> ---
Confirmed.  Here's a creduced test case. This fails even with -O1 -mcpu=power8:

bergner@pike:~/gcc/BUGS/PR103317$ cat bug.i
int a, b;
char c;
void
d (void)
{
  char e = c;
  if (b)
    if (c < 16 - 11)
      e = 16 - 11;
  if (e > 8)
    e = 8;
  a = e;
}

bergner@pike:~/gcc/BUGS/PR103317$ gcc -S -O1 -mcpu=power8 bug.i 
bug.i: In function ‘d’:
bug.i:13:1: error: definition in block 3 does not dominate use in block 2
   13 | }
      | ^
for SSA_NAME: _2 in statement:
e_5 = PHI <_2(2), _4(3)>
PHI argument
_2
for PHI node
e_5 = PHI <_2(2), _4(3)>
during GIMPLE pass: phiopt
bug.i:13:1: internal compiler error: verify_ssa failed
0x11989b73 verify_ssa(bool, bool)

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #3 from Peter Bergner <bergner at gcc dot gnu.org> ---
A not too old trunk build didn't ICE, so this looks new.  I'll bisect it to
find the bad commit.

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

Peter Bergner <bergner at gcc dot gnu.org> changed:

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

--- Comment #4 from Peter Bergner <bergner at gcc dot gnu.org> ---
My bisect flagged the following commit as causing the ICE:

commit f98f373dd822b35c52356b753d528924e9f89678
Commit:     Richard Biener <rguenther@suse.de>
CommitDate: Tue Nov 16 11:31:05 2021 +0100

    tree-optimization/102880 - make PHI-OPT recognize more CFGs
----
    This allows extra edges into the middle BB for the PHI-OPT
    transforms using replace_phi_edge_with_variable that do not
    end up moving stmts from that middle BB.  This avoids regressing
    gcc.dg/tree-ssa/ssa-hoist-4.c with the actual fix for PR102880
    where CFG cleanup has the choice to remove two forwarders and
    picks "the wrong" leading to

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Hmm, I fixed one like this yesterday. Are you sure it is not fixed?

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
https://gcc.gnu.org/pipermail/gcc-cvs/2021-November/356905.html

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #7 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #5)
> Hmm, I fixed one like this yesterday. Are you sure it is not fixed?

Yes, I built with a trunk from this morning and just verified it still ICEs
with a new checkout (6f4ac4f81f89caac7e74127ed2e6db6bbb3d7426).

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I will go look into this after a nap.

[Bug tree-optimization/103317] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |pinskia at gcc dot gnu.org
             Status|NEW                         |ASSIGNED
   Target Milestone|---                         |12.0

[Bug tree-optimization/103317] [12 regression] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #9 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Created attachment 51834
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51834&action=edit
patch which I am testing

This is the patch which I am testing which should get us back to where we were.

[Bug tree-optimization/103317] [12 regression] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #10 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #9)
> Created attachment 51834 [details]
> patch which I am testing
> 
> This is the patch which I am testing which should get us back to where we
> were.

I can confirm this fixes the ICE on the reduced test case as well as the
original large test case.  Thanks for looking into this Andrew!

[Bug tree-optimization/103317] [12 regression] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch
                URL|                            |https://gcc.gnu.org/piperma
                   |                            |il/gcc-patches/2021-Novembe
                   |                            |r/584926.html

--- Comment #11 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Patch submitted:
https://gcc.gnu.org/pipermail/gcc-patches/2021-November/584926.html

[Bug tree-optimization/103317] [12 regression] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

--- Comment #12 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Andrew Pinski <pinskia@gcc.gnu.org>:

https://gcc.gnu.org/g:09d462146b3107c665265b11ad925c61a91c6efb

commit r12-5393-g09d462146b3107c665265b11ad925c61a91c6efb
Author: Andrew Pinski <apinski@marvell.com>
Date:   Thu Nov 18 23:38:30 2021 +0000

    Fix PR 103317, ICE after PHI-OPT, minmax_replacement producing invalid SSA

    The problem is r12-5300-gf98f373dd822b35c allows phiopt to recognize more
basic blocks
    but missed one location where the basic block does not need to be empty but
still
    needs to have a single predecessor. This patch fixes that over sight.

    OK? Bootstrapped and tested on x86_64-linux-gnu with no regressions.

            PR tree-optimization/103317

    gcc/ChangeLog:

            * tree-ssa-phiopt.c (minmax_replacement): For the non empty
            middle bb case, check to make sure it has a single predecessor.

    gcc/testsuite/ChangeLog:

            * gcc.c-torture/compile/pr103317-1.c: New test.

[Bug tree-optimization/103317] [12 regression] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

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

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

--- Comment #13 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Fixed. thanks for the reduced testcase even.

[Bug tree-optimization/103317] [12 regression] Spec 2017 benchmark blender_r fails with -Ofast on PowerPc (power9, power10)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103317

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

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

--- Comment #14 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 103321 has been marked as a duplicate of this bug. ***