[Bug tree-optimization/98602] New: Failure to optimise vector “x > -100 ? x : -100” to MAX_EXPR

[Bug tree-optimization/98602] New: Failure to optimise vector “x > -100 ? x : -100” to MAX_EXPR

[rsandifo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98602

            Bug ID: 98602
           Summary: Failure to optimise vector “x > -100 ? x : -100” to
                    MAX_EXPR
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: rsandifo at gcc dot gnu.org
  Target Milestone: ---

We don't fold:

  typedef int v4si __attribute__ ((vector_size(16)));
  v4si f (v4si x) { return x > -100 ? x : -100; }

to a MAX_EXPR, even though we do if -100 is replaced with a
nonnegative constant or a variable.  This may be related to
the remaining part of PR95906.

[Bug tree-optimization/98602] Failure to optimise vector “x > -100 ? x : -100” to MAX_EXPR

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98602

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-01-08

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Guess we miss folders for this since it is represented as VEC_COND_EXPR from
the start (the testcase only builds with the C++ frontend) and scalar MIN/MAX
are usually detected by phiopt.  Ah, fold_cond_expr_with_comparison doesn't
handle adjusted constants:

return <retval> =  VEC_COND_EXPR < x >= { -99, -99, -99, -99 } , x , { -100,
-100, -100, -100 } > ;

[Bug tree-optimization/98602] Failure to optimise vector “x > -100 ? x : -100” to MAX_EXPR

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98602

--- Comment #2 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Richard Sandiford <rsandifo@gcc.gnu.org>:

https://gcc.gnu.org/g:48c7f5b88116c894d76209db0f47216ed5ccd333

commit r11-6595-g48c7f5b88116c894d76209db0f47216ed5ccd333
Author: Richard Sandiford <richard.sandiford@arm.com>
Date:   Mon Jan 11 18:03:21 2021 +0000

    aarch64: Add support for unpacked SVE mult, max and min

    This patch makes the SVE_INT_BINARY_IMM patterns support
    unpacked arithmetic, covering MUL, SMAX, SMIN, UMAX and UMIN.
    For min and max, the type suffix must be taken from the element
    size rather than the container size.

    The XFAILs are due to PR98602.

    gcc/
            * config/aarch64/aarch64-sve.md (<SVE_INT_BINARY_IMM:optab><mode>3)
            (@aarch64_pred_<SVE_INT_BINARY_IMM:optab><mode>)
            (*post_ra_<SVE_INT_BINARY_IMM:optab><mode>3): Extend from
SVE_FULL_I
            to SVE_I.

    gcc/testsuite/
            PR testsuite/98602
            * g++.target/aarch64/sve/max_1.C: New test.
            * g++.target/aarch64/sve/min_1.C: Likewise.
            * gcc.target/aarch64/sve/mul_2.c: Likewise.

[Bug tree-optimization/98602] Failure to optimise vector “x > -100 ? x : -100” to MAX_EXPR

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98602

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
           Severity|normal                      |enhancement
                 CC|                            |pinskia at gcc dot gnu.org
             Status|NEW                         |ASSIGNED

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I might get to this one when I move minmax_replacement from phiopt to match.pd.

[Bug tree-optimization/98602] Failure to optimise vector “x > -100 ? x : -100” to MAX_EXPR

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98602

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

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

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Actually I am not going to handle this as the important parts of
fold_cond_expr_with_comparison have been ported already to match.pd.
But need extending for vector cases.