[Bug middle-end/118801] Excessive compile time with -g -O2 -fpeel-loops -fno-var-tracking

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118801

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2025-02-10
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot 
gnu.org
     Ever confirmed|0                           |1
                 CC|                            |aoliva at gcc dot gnu.org,
                   |                            |jakub at gcc dot gnu.org

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  The compile-time is almost entirely because of (useless) debug
stmts, in particular # DEBUG BEGIN_STMT.

  [runData/keep/in.6699.c:450:13] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:451:13] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:452:13] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:452:40 discrim 1] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:454:17] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:455:17] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:456:17] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:457:17] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:458:17] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:459:17] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:460:17] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:461:17] # DEBUG BEGIN_STMT
  [runData/keep/in.6699.c:462:17] # DEBUG BEGIN_STMT

we're pruning "redunant" DEBUG_BINDs in DCE already, but do nothing to
DEBUG BEGIN STMT.

I _think_ we can just keep the last of a series of DEBUG BEGIN STMT
(what about intermediate DEBUG BINDs?)?

Out of 1982391 lines in t.c.269t.optimized this testcase has 1981380
DEBUG BEGIN_STMT.  I'd call it 100% with some rounding error.