[Bug tree-optimization/101991] New: bit_and or bit_ior with an invariant inside loop is not pulled out of the loop

[Bug tree-optimization/101991] New: bit_and or bit_ior with an invariant inside loop is not pulled out of the loop

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101991

            Bug ID: 101991
           Summary: bit_and or bit_ior with an invariant inside loop is
                    not pulled out of the loop
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: enhancement
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: pinskia at gcc dot gnu.org
  Target Milestone: ---

Take these two functions:
int f(int t, int d, int e)
{
  int r = d;
  for(int i = 0; i < t; i++)
    r &= e;
  return r;
}

int f1(int t, int d, int e)
{
  int r = d;
  if (0 < t)
    r &= e;
  return r;
}
They should produce the same code.  Right now f still has a loop in it.
The same is true with bit_ior.

Note the code is really bad when the vectorizer comes around really.
I forgot how I found this either.
Note clang/ICC also vectorize this code crappily.
clang on the trunk almost gets there but still has a loop:

        movl    %esi, %eax
        xorl    %esi, %esi
        testl   %edi, %edi
        cmovgl  %edi, %esi
        jle     .LBB0_6
# %bb.1:
        leal    -1(%rsi), %edi
        movl    %esi, %ecx
        andl    $7, %ecx
        cmpl    $7, %edi
        jb      .LBB0_4
# %bb.2:
        andl    $-8, %esi
        negl    %esi
        .p2align        4, 0x90
.LBB0_3:                                # =>This Inner Loop Header: Depth=1
        addl    $8, %esi
        jne     .LBB0_3
.LBB0_4:
        andl    %edx, %eax
        testl   %ecx, %ecx
        je      .LBB0_6
        .p2align        4, 0x90
.LBB0_5:                                # =>This Inner Loop Header: Depth=1
        addl    $-1, %ecx
        jne     .LBB0_5
.LBB0_6:
        retq

[Bug tree-optimization/101991] bit_and or bit_ior with an invariant inside loop is not pulled out of the loop

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101991

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-08-20

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
  <bb 3> [local count: 955630225]:
  # r_11 = PHI <r_8(6), d_3(D)(5)>
  r_8 = e_7(D) & r_11;

I wonder if this is sth for phiopt to pattern match.  In principle VN
would need to figure (for PRE) that the PHI translated d_3(D) & e_7(D)
is equal to r_8.  So the "trick" (aka pattern-matching) could be done
during phi_translation.

But then both look like a hack.  Curiously when we do

int f(int t, int d, int e)
{
  int r = d & e;
  for(int i = 0; i < t; i++)
    r &= e;
  return r;
}

aka "peel" one iteration, then CCP is what eliminates the in-loop AND.
Ah, that's because we simplify d & e & e since we optimistically start
with just the entry edge value.  And that remains so.  With PRE
we're not fully re-doing VN of PHIs but phi-translation seeks to
re-use existing value-numbers where possible, so a programmatic approach
doesn't work here.

[Bug tree-optimization/101991] bit_and or bit_ior with an invariant inside loop is not pulled out of the loop

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101991

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 105735 has been marked as a duplicate of this bug. ***

[Bug tree-optimization/101991] bit_and or bit_ior with an invariant inside loop is not pulled out of the loop

[lingling.kong7 at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101991

kong lingling <lingling.kong7 at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |lingling.kong7 at gmail dot com

--- Comment #3 from kong lingling <lingling.kong7 at gmail dot com> ---
Fixed in gcc 13.0.

[Bug tree-optimization/101991] bit_and or bit_ior with an invariant inside loop is not pulled out of the loop

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101991

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|NEW                         |RESOLVED
   Target Milestone|---                         |13.0

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Fixed.