[Bug c/100473] New: O2 vs optimization flags

[Bug c/100473] New: O2 vs optimization flags

[stanislav.simek at siemens dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100473

            Bug ID: 100473
           Summary: O2 vs optimization flags
           Product: gcc
           Version: 10.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: stanislav.simek at siemens dot com
  Target Milestone: ---

Hi all,
May I ask you, why the object file created with option -O2 is different than
with all flags (for O1 + O2) mentioned here:
https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html


Eq.
Why object generated with -O2 is different (in assembly)

than object generated with:

=== O1 part===

-fauto-inc-dec 
-fbranch-count-reg 
-fcombine-stack-adjustments 
-fcompare-elim 
-fcprop-registers 
-fdce 
-fdefer-pop 
-fdelayed-branch 
-fdse 
-fforward-propagate 
-fguess-branch-probability 
-fif-conversion 
-fif-conversion2 
-finline-functions-called-once 
-fipa-modref 
-fipa-profile 
-fipa-pure-const 
-fipa-reference 
-fipa-reference-addressable 
-fmerge-constants 
-fmove-loop-invariants 
-fomit-frame-pointer 
-freorder-blocks 
-fshrink-wrap 
-fshrink-wrap-separate 
-fsplit-wide-types 
-fssa-backprop 
-fssa-phiopt 
-ftree-bit-ccp 
-ftree-ccp 
-ftree-ch 
-ftree-coalesce-vars 
-ftree-copy-prop 
-ftree-dce 
-ftree-dominator-opts 
-ftree-dse 
-ftree-forwprop 
-ftree-fre 
-ftree-phiprop 
-ftree-pta 
-ftree-scev-cprop 
-ftree-sink 
-ftree-slsr 
-ftree-sra 
-ftree-ter 
-funit-at-a-time

 === O2 part === 

-falign-functions  -falign-jumps 
-falign-labels  -falign-loops 
-fcaller-saves 
-fcode-hoisting 
-fcrossjumping 
-fcse-follow-jumps  -fcse-skip-blocks 
-fdelete-null-pointer-checks 
-fdevirtualize  -fdevirtualize-speculatively 
-fexpensive-optimizations 
-ffinite-loops 
-fgcse  -fgcse-lm  
-fhoist-adjacent-loads 
-finline-functions 
-finline-small-functions 
-findirect-inlining 
-fipa-bit-cp  -fipa-cp  -fipa-icf 
-fipa-ra  -fipa-sra  -fipa-vrp 
-fisolate-erroneous-paths-dereference 
-flra-remat 
-foptimize-sibling-calls 
-foptimize-strlen 
-fpartial-inlining 
-fpeephole2 
-freorder-blocks-algorithm=stc 
-freorder-blocks-and-partition  -freorder-functions 
-frerun-cse-after-loop  
-fschedule-insns  -fschedule-insns2 
-fsched-interblock  -fsched-spec 
-fstore-merging 
-fstrict-aliasing 
-fthread-jumps 
-ftree-builtin-call-dce 
-ftree-pre 
-ftree-switch-conversion  -ftree-tail-merge 
-ftree-vrp

[Bug c/100473] O2 vs optimization flags

[schwab@linux-m68k.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100473

--- Comment #1 from Andreas Schwab <schwab@linux-m68k.org> ---
$ git grep 'optimize >= 2'
gcc/ChangeLog-2008:     100 for optimize >= 2.
gcc/ChangeLog-2010:     only conditionally on optimize >= 2.
gcc/ada/gcc-interface/trans.c:        && optimize >= 2
gcc/c-family/c-opts.c:                 optimize >= 2 && cxx_dialect >= cxx11);
gcc/config/aarch64/aarch64-cc-fusion.cc:  virtual bool gate (function *) {
return TARGET_SVE && optimize >= 2; }
gcc/config/aarch64/cortex-a57-fma-steering.c:         && optimize >= 2;
gcc/config/aarch64/falkor-tag-collision-avoidance.c:          && optimize >=
2);
gcc/config/arm/arm.c:      else if (optimize >= 2)
gcc/config/arm/arm.h:                           ((optimize >= 2) \
gcc/config/msp430/msp430.c:  bool speed = (!optimize_size && optimize >= 2);
gcc/config/msp430/msp430.c:           && (optimize >= 2 && !optimize_size))))
gcc/config/msp430/msp430.c:           && (optimize >= 2 && !optimize_size)
gcc/config/msp430/msp430.h:#define NO_FUNCTION_CSE (optimize >= 2 &&
!optimize_size)
gcc/expr.c:      && optimize >= 2
gcc/expr.c:      if (SCALAR_INT_MODE_P (mode) && optimize >= 2)
gcc/opts.c:  opt2 = (opts->x_optimize >= 2);
gcc/tree-ssa-dce.c:  return perform_tree_ssa_dce (/*aggressive=*/optimize >=
2);
gcc/tree-ssa-loop-ivcanon.c:  virtual bool gate (function *) { return optimize
>= 2; }

[Bug c/100473] O2 vs optimization flags

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100473

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
                 CC|                            |jakub at gcc dot gnu.org
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
And
git grep 'optimize > 1'
gcc/config/arc/arc.c:    return (optimize > 1 && !TARGET_NO_COND_EXEC);
gcc/config/arc/arc.c:      if (optimize > 1 && !TARGET_NO_COND_EXEC)
gcc/config/i386/i386-expand.c:     && optimize > 1)
gcc/config/i386/i386-features.c:              && TARGET_STV && TARGET_SSE2 &&
optimize > 1);
gcc/config/mcore/mcore.c:  if (optimize > 1 && blk_size > 2)
gcc/config/rl78/rl78.c:  if (optimize > 1)
gcc/cse.c:  else if (tem == 1 || optimize > 1)
gcc/dce.c:      return optimize > 1 && flag_dce && dbg_cnt (dce_ud);
gcc/df-core.c:  if (optimize > 1)
gcc/early-remat.c:    return optimize > 1 && NUM_POLY_INT_COEFFS > 1;
gcc/ira.c:  if (optimize > 1)
gcc/ira.c:  if (optimize > 1)
gcc/tree-eh.c:  if (optimize > 1)
gcc/tree-ssa-sccvn.c:      return flag_tree_fre != 0 && (may_iterate ||
optimize > 1);
gcc/tree-ssa-sccvn.c:  bool iterate_p = may_iterate && (optimize > 1);
git grep 'optimize <= 1'
gcc/df-problems.c:  df_live->optional_p = (optimize <= 1);
git grep 'optimize < 2'
gcc/builtins.c:  if (optimize < 2 || optimize_insn_for_size_p ())
gcc/cfgexpand.c:  if (toplevel && optimize < 2)
gcc/config/arc/arc.c:  if (optimize < 2 || TARGET_NO_COND_EXEC)
gcc/config/pa/pa.c:  if (optimize < 2)
gcc/dse.c:  if (optimize < 2)
gcc/expr.c:  if (optimize < 2)
gcc/loop-invariant.c:      if (optimize < 2)
gcc/tree-ssa-loop-im.c:  if (optimize < 2)

I think we've never claimed that -O2 stands for -O1 + specific options,
similarly to how -O1 certainly is not -O0 + lots of options (even much more
different, because if (optimize) is very common and is used in gating tests of
whole pass sets.

[Bug middle-end/100473] O2 vs optimization flags

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100473

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c                           |middle-end

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #2) 
> I think we've never claimed that -O2 stands for -O1 + specific options,
> similarly to how -O1 certainly is not -O0 + lots of options (even much more
> different, because if (optimize) is very common and is used in gating tests
> of whole pass sets.

From
https://gcc.gnu.org/onlinedocs/gcc-11.1.0/gcc/Optimize-Options.html#Optimize-Options
:
Not all optimizations are controlled directly by a flag. Only optimizations
that have a flag are listed in this section.

[Bug middle-end/100473] O2 vs optimization flags

[stanislav.simek at siemens dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100473

--- Comment #4 from Stanislav Šimek <stanislav.simek at siemens dot com> ---
Okay, thank you a lot for explanation.