[Bug fortran/64530] New: [4.9 Regression] Incorrect calculation when assigning to array with -O3

[Bug fortran/64530] New: [4.9 Regression] Incorrect calculation when assigning to array with -O3

[jkaus at ucsd dot edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

            Bug ID: 64530
           Summary: [4.9 Regression] Incorrect calculation when assigning
                    to array with -O3
           Product: gcc
           Version: 4.9.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: jkaus at ucsd dot edu

Created attachment 34398
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=34398&action=edit
Minimal example illustrating the issue.

With gfortran 4.9 using -O3, the assignment to an array
can occur before the rest of the calculation, giving
an incorrect result. I have attached a minimal example
which shows the bug along with modified code that 
produces the correct results.

With O3 we get the wrong result (0.0):
gfortran -O3 a.F90
./a.out
> buggy   0.0000000000000000        0.0000000000000000        0.0000000000000000     
> works   3.5000000000000000        3.5000000000000000        3.5000000000000000  

With O2 we get the correct result (3.5):
gfortran -O2 a.F90
./a.out 
> buggy   3.5000000000000000        3.5000000000000000        3.5000000000000000     
> works   3.5000000000000000        3.5000000000000000        3.500000000000000

gfortran -v
Using built-in specs.
COLLECT_GCC=gfortran
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Ubuntu
4.9.2-0ubuntu1~12.04' --with-bugurl=file:///usr/share/doc/gcc-4.9/README.Bugs
--enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr
--program-suffix=-4.9 --enable-shared --enable-linker-build-id
--libexecdir=/usr/lib --without-included-gettext --enable-threads=posix
--with-gxx-include-dir=/usr/include/c++/4.9 --libdir=/usr/lib --enable-nls
--with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug
--enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-vtable-verify
--enable-plugin --with-system-zlib --disable-browser-plugin
--enable-java-awt=gtk --enable-gtk-cairo
--with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64/jre --enable-java-home
--with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64
--with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.9-amd64
--with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar
--enable-objc-gc --enable-multiarch --disable-werror --with-arch-32=i686
--with-abi=m64 --with-multilib-list=m32,m64 --enable-multilib
--with-tune=generic --enable-checking=release --build=x86_64-linux-gnu
--host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.9.2 (Ubuntu 4.9.2-0ubuntu1~12.04)

[Bug fortran/64530] [4.9 Regression] Incorrect calculation when assigning to array with -O3

[anlauf at gmx dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

Harald Anlauf <anlauf at gmx dot de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |anlauf at gmx dot de

--- Comment #1 from Harald Anlauf <anlauf at gmx dot de> ---
(In reply to jkaus from comment #0)
> With O3 we get the wrong result (0.0):
> gfortran -O3 a.F90
> ./a.out
> > buggy   0.0000000000000000        0.0000000000000000        0.0000000000000000     
> > works   3.5000000000000000        3.5000000000000000        3.5000000000000000  

Interesting bug.  I can confirm that it does not occur with 4.8,
but with 4.9 and 5.

Note that if one interchanges in subroutine "buggy" the lines

    old(i) = ave(i)
    tmp(i) = tmp(i) / scale_factor

to

    tmp(i) = tmp(i) / scale_factor
    old(i) = ave(i)

I also get the correct result.

With gfortran 4.9 and 5 I also get the correct result with -O3
when e.g. using UBSAN (-fsanitize=undefined); it probably also
reduces the optimization.

Since the dump tree looks the same for 4.8 and 4.9, could it be
a middle-end bug?

[Bug fortran/64530] [4.9 Regression] Incorrect calculation when assigning to array with -O3

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-01-07
                 CC|                            |rguenth at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #2 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
(1) The problem goes away if the code is compiled with '-O3
-fno-tree-loop-vectorize'.
(2) The bug appeared between r204000 (OK, 2013-10-24) and r204945 (buggy,
2013-11-18).

[Bug tree-optimization/64530] [4.9 Regression] Incorrect calculation when assigning to array with -O3

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
          Component|fortran                     |tree-optimization
      Known to work|                            |4.8.5
      Known to fail|                            |4.9.2, 5.0

--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> Since the dump tree looks the same for 4.8 and 4.9, could it be
> a middle-end bug?

Indeed! reassigned to tree optimization.

[Bug tree-optimization/64530] [4.9/5 Regression] Incorrect calculation when assigning to array with -O3

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org
   Target Milestone|---                         |4.9.3

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r204062.

[Bug tree-optimization/64530] [4.9/5 Regression] Incorrect calculation when assigning to array with -O3

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
Mine.

[Bug tree-optimization/64530] [4.9/5 Regression] Incorrect calculation when assigning to array with -O3

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
Author: rguenth
Date: Mon Jan 12 15:37:07 2015
New Revision: 219474

URL: https://gcc.gnu.org/viewcvs?rev=219474&root=gcc&view=rev
Log:
2015-01-12  Richard Biener  <rguenther@suse.de>

    PR tree-optimization/64530
    * tree-loop-distribution.c (pg_add_dependence_edges): Shuffle
    back dr1.

    * gfortran.dg/pr64530.f90: New testcase.

Added:
    trunk/gcc/testsuite/gfortran.dg/pr64530.f90
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-loop-distribution.c

[Bug tree-optimization/64530] [4.9 Regression] Incorrect calculation when assigning to array with -O3

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |5.0
            Summary|[4.9/5 Regression]          |[4.9 Regression] Incorrect
                   |Incorrect calculation when  |calculation when assigning
                   |assigning to array with -O3 |to array with -O3
      Known to fail|5.0                         |

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed on trunk sofar.

[Bug tree-optimization/64530] [4.9 Regression] Incorrect calculation when assigning to array with -O3

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
Author: rguenth
Date: Fri Feb 20 07:32:08 2015
New Revision: 220835

URL: https://gcc.gnu.org/viewcvs?rev=220835&root=gcc&view=rev
Log:
2015-02-20  Richard Biener  <rguenther@suse.de>

    Backport from mainline
    2015-01-12  Richard Biener  <rguenther@suse.de>

    PR tree-optimization/64530
    * tree-loop-distribution.c (pg_add_dependence_edges): Shuffle
    back dr1.

    * gfortran.dg/pr64530.f90: New testcase.

    2015-02-13  Richard Biener  <rguenther@suse.de>

    PR lto/64373
    * lto-streamer-out.c (tree_is_indexable): Guard for NULL
    DECL_CONTEXT.

    * gcc.dg/lto/pr64373_0.c: New testcase.

    2015-02-16  Richard Biener  <rguenther@suse.de>

    PR tree-optimization/63593
    * tree-predcom.c (execute_pred_commoning_chain): Delay removing
    stmts and releasing SSA names until...
    (execute_pred_commoning): ... after processing all chains.

    * gcc.dg/pr63593.c: New testcase.

    2015-02-18  Richard Biener  <rguenther@suse.de>

    PR tree-optimization/65063
    * tree-predcom.c (determine_unroll_factor): Return 1 if we
    have replaced looparound PHIs.

    * gcc.dg/pr65063.c: New testcase.

Added:
    branches/gcc-4_9-branch/gcc/testsuite/gcc.dg/lto/pr64373_0.c
    branches/gcc-4_9-branch/gcc/testsuite/gcc.dg/pr63593.c
    branches/gcc-4_9-branch/gcc/testsuite/gcc.dg/pr65063.c
    branches/gcc-4_9-branch/gcc/testsuite/gfortran.dg/pr64530.f90
Modified:
    branches/gcc-4_9-branch/gcc/ChangeLog
    branches/gcc-4_9-branch/gcc/lto-streamer-out.c
    branches/gcc-4_9-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_9-branch/gcc/tree-loop-distribution.c
    branches/gcc-4_9-branch/gcc/tree-predcom.c

[Bug tree-optimization/64530] [4.9 Regression] Incorrect calculation when assigning to array with -O3

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64530

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
      Known to work|                            |4.9.3
         Resolution|---                         |FIXED

--- Comment #9 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed.