[Bug ipa/64425] New: [5 Regression] r219076 causes 36% tramp3d slowdown

[Bug ipa/64425] New: [5 Regression] r219076 causes 36% tramp3d slowdown

[trippels at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64425

            Bug ID: 64425
           Summary: [5 Regression] r219076 causes 36% tramp3d slowdown
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: ipa
          Assignee: unassigned at gcc dot gnu.org
          Reporter: trippels at gcc dot gnu.org
                CC: hubicka at gcc dot gnu.org

Seen on ppc64. Before r219076:

 % g++ -Ofast -w tramp3d-v4.cpp
 % ./a.out --cartvis 1.0 0.0 --rhomin 1e-8 -n 20
Time spent in iteration: 2.40742

After r219076:

 % g++ -Ofast -w tramp3d-v4.cpp
 % ./a.out --cartvis 1.0 0.0 --rhomin 1e-8 -n 20
Time spent in iteration: 3.77864

[Bug ipa/64425] [5 Regression] r219076 causes 36% tramp3d slowdown

[trippels at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64425

--- Comment #1 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
On my x86_64 machine it's worse: 2.93392 vs. 5.84974.

[Bug ipa/64425] [5 Regression] r219076 causes 36% tramp3d slowdown

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64425

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hjl.tools at gmail dot com
   Target Milestone|---                         |5.0

--- Comment #2 from H.J. Lu <hjl.tools at gmail dot com> ---
There is also a correctness issue.  On Linux/x32, r219076 caused:

FAIL: 23_containers/unordered_set/insert/hash_policy.cc execution test
FAIL: 23_containers/unordered_set/max_load_factor/robustness.cc execution test

[Bug ipa/64425] [5 Regression] r219076 causes 36% tramp3d slowdown

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64425

--- Comment #3 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Yep, I looked into this before comitting the patch. I have tracked it down to
inlining of function UniformRectilinearMesh<MeshTraits<3, double,
UniformRectilinearTag, CartesianTag, 3> >::cellPosition(Loc<3> const&) const:

  inline PointType_t cellPosition(const Loc_t &loc) const                       
    {                                                                           
      PointType_t point;                                                        
      for (int i=0; i<dimensions; i++)                                          
 point(i) = origin()(i) + spacings()(i)                                         
   * (loc[i].first() - physicalCellDomain()[i].first() + 0.5);                  
      return point;                                                             
    }                                                                           

this function is somewhat evil: dimensons is a small compile time constant and
the loop unrolls completely into quite simple code.  This is however not done
prior inlining.  Inliner thus think that the function is 168 instructions and
the speedup for inlining accounts only removing some of this-> references,
overall 9%.

As an effect of various roundoff errors being eliminated by the sreal code, the
functions sinks down in the priority queue quite a bit bellow the
inline-unit-growth cutoff.  There seems to be no issues with priority
calculation, just the functions looks bit uncool for inlining and got lucky in
the old model (that was for years tested on tramp3d)

The patch I sent for simplification of priority queue solves the failure:
https://gcc.gnu.org/ml/gcc-patches/2015-01/msg00004.html

The patch seem win in general on benchmarks I tested (it is big win for tramp3d
so I can reduce unit growth to 15%) I however would like first the testers to
get restarted after vacations and there is one regression I want to analyze
first in my testing.

[Bug ipa/64425] [5 Regression] r219076 causes 36% tramp3d slowdown

[trippels at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64425

--- Comment #4 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
I've tested the patch, too. 
There were two issues.

1) on ppc64:

(after training)
trippels@gcc2-power8 ~ % ~/gcc_test/usr/local/bin/g++ -Ofast -fprofile-use -w
tramp3d-v4.cpp
tramp3d-v4.cpp:64206:1: internal compiler error: in edge_badness, at
ipa-inline.c:1030
 }
 ^
0x110330ff edge_badness  
        ../../gcc/gcc/ipa-inline.c:1030
0x11039083 update_edge_key
        ../../gcc/gcc/ipa-inline.c:1059
0x11039083 inline_small_functions
        ../../gcc/gcc/ipa-inline.c:1634
0x1103b0af ipa_inline
        ../../gcc/gcc/ipa-inline.c:2175
0x1103b0af execute
        ../../gcc/gcc/ipa-inline.c:2548
Please submit a full bug report,
with preprocessed source if appropriate.

2) Firefox PGO/LTO build on x86_64 segfaults on start.

[Bug ipa/64425] [5 Regression] r219076 causes 36% tramp3d slowdown

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64425

--- Comment #5 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Author: hubicka
Date: Mon Jan 12 09:28:15 2015
New Revision: 219452

URL: https://gcc.gnu.org/viewcvs?rev=219452&root=gcc&view=rev
Log:

    PR ipa/63967
    PR ipa/64425
    * ipa-inline.c (compute_uninlined_call_time,
    compute_inlined_call_time): Use counts for extra precision when
    needed possible.
    (big_speedup_p): Fix formating.
    (RELATIVE_TIME_BENEFIT_RANGE): Remove.
    (relative_time_benefit): Remove.
    (edge_badness): Turn DECL_DISREGARD_INLINE_LIMITS into hint;
    merge guessed and read profile paths.
    (inline_small_functions): Count only !optimize_size functions into
    initial size; be more lax about sanity check when profile is used;
    be sure to update inlined function profile when profile is read.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-inline.c

[Bug ipa/64425] [5 Regression] r219076 causes 36% tramp3d slowdown

[trippels at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64425

Markus Trippelsdorf <trippels at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED

--- Comment #6 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
Fixed. Thanks. Even -flto is fine now.