[Bug ipa/99834] New: missed optimization for dead code elimination at -O3 (vs. -O2)

[Bug ipa/99834] New: missed optimization for dead code elimination at -O3 (vs. -O2)

[zhendong.su at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99834

            Bug ID: 99834
           Summary: missed optimization for dead code elimination at -O3
                    (vs. -O2)
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: ipa
          Assignee: unassigned at gcc dot gnu.org
          Reporter: zhendong.su at inf dot ethz.ch
                CC: marxin at gcc dot gnu.org
  Target Milestone: ---

[590] % gcctk -v
Using built-in specs.
COLLECT_GCC=gcctk
COLLECT_LTO_WRAPPER=/local/suz-local/software/local/gcc-trunk/libexec/gcc/x86_64-pc-linux-gnu/11.0.1/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: ../gcc-trunk/configure --disable-bootstrap
--prefix=/local/suz-local/software/local/gcc-trunk --enable-languages=c,c++
--disable-werror --enable-multilib --with-system-zlib
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 11.0.1 20210330 (experimental) [master revision
8aac913adfc:ff4ebc2f17c:65374af219f9c5c594951a07e766fe70c1136a1f] (GCC) 
[591] % 
[591] % gcctk -O2 -S -o O2.s small.c
[592] % gcctk -O3 -S -o O3.s small.c
[593] % 
[593] % wc O2.s O3.s
  43   90  704 O2.s
  61  126  950 O3.s
 104  216 1654 total
[594] % 
[594] % grep foo O2.s
[595] % grep foo O3.s
        call    foo
        call    foo
[596] % 
[596] % cat small.c
extern void foo(void);
static int a, b, c;
static int d() {
  for (a = 0; a < 1; a++)
    b = 1;
  for (; b; b--)
    for (; c; c--)
      ;
  return 0;
}
static void e() {
  if (d())
    foo();
}
int main() {
  e();
  e();
  return 0;
}

[Bug ipa/99834] missed optimization for dead code elimination at -O3 (vs. -O2)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99834

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2021-03-31
            Version|unknown                     |11.0
                 CC|                            |hubicka at gcc dot gnu.org
           Keywords|                            |missed-optimization
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Similar to the other bug - when 'main' is named 'bar' then we optimize it OK at
-O3, the difference being again loop header copying not applied in "cold"
context
but inlining is.

Not sure if we want to call this a bug, but then the inconsistency is odd.

[Bug ipa/99834] missed optimization for dead code elimination at -O3 (vs. -O2)

[zhendong.su at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99834

--- Comment #2 from Zhendong Su <zhendong.su at inf dot ethz.ch> ---
Hi Richard and all, thanks for analyzing these reports! 

I have some more cases, and wonder whether you folks would prefer that I open a
meta issue report and append these (and others that we find) to that report
(rather than filing these one by one). Thanks.

[Bug ipa/99834] missed optimization for dead code elimination at -O3 (vs. -O2)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99834

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The other way of optimizing this is to see that the return value of d is 0
(even early on) and do a const prop of that ... IN a sense an IPA-CCP for
return values rather than for argument passing.