[Bug rtl-optimization/68128] New: A huge regression in Parboil v2.5 OpenMP CUTCP test (2.5 times lower performance)

[Bug rtl-optimization/68128] New: A huge regression in Parboil v2.5 OpenMP CUTCP test (2.5 times lower performance)

[t.artem at mailcity dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68128

            Bug ID: 68128
           Summary: A huge regression in Parboil v2.5 OpenMP CUTCP test
                    (2.5 times lower performance)
           Product: gcc
           Version: 5.3.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: t.artem at mailcity dot com
  Target Milestone: ---

According to Phoronix,
http://www.phoronix.com/scan.php?page=article&item=fedora-2023-bench&num=4 ,
for Parboil v2.5 OpenMP CUTCP test GCC in Fedora 22/23 produces the code which
runs 2.5 times slower than the code produced by Fedora 20.

Fedora 20: GCC 4.8.2 (good performance)
Fedora 21: GCC 4.9.1 (good performance)
Fedora 22: GCC 5.1.1 with patches (huge regression)
Fedora 23: GCC 5.1.1 with even more patches (huge regression)

[Bug rtl-optimization/68128] A huge regression in Parboil v2.5 OpenMP CUTCP test (2.5 times lower performance)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68128

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |missed-optimization, openmp
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Not very much information (compile flags?)

[Bug rtl-optimization/68128] A huge regression in Parboil v2.5 OpenMP CUTCP test (2.5 times lower performance)

[t.artem at mailcity dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68128

--- Comment #2 from Artem S. Tashkinov <t.artem at mailcity dot com> ---
(In reply to Richard Biener from comment #1)
> Not very much information (compile flags?)

CPU: Intel Xeon E5-2687W v3 (
http://ark.intel.com/products/81909/Intel-Xeon-Processor-E5-2687W-v3-25M-Cache-3_10-GHz
)

GCC Flags: (CXX) g++ options: -lm -lpthread -lgomp -ffast-math -fopenmp

Something tells me some other GCC flags were used but the benchmarker hasn't
yet replied to my question.