[Bug testsuite/59494] New: [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[Bug testsuite/59494] New: [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

            Bug ID: 59494
           Summary: [4.9 Regression] FAIL:
                    gfortran.dg/vect/fast-math-mgrid-resid.f
                    scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: testsuite
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dominiq at lps dot ens.fr
                CC: hjl.tools at gmail dot com, howarth at bromo dot med.uc.edu,
                    jakub at gcc dot gnu.org

The test gfortran.dg/vect/fast-math-mgrid-resid.f fails after r203979 if gcc is
configured with --with-cpu=core* (default for darwin). This is silenced if
-mtune=generic is added to the dg-options list:

--- gcc/testsuite/gfortran.dg/vect/fast-math-mgrid-resid.f    2013-04-08
15:53:21.000000000 +0200
+++ ../work/gcc/testsuite/gfortran.dg/vect/fast-math-mgrid-resid.f   
2013-12-12 21:51:53.000000000 +0100
@@ -1,7 +1,7 @@
 ! { dg-do compile { target i?86-*-* x86_64-*-* } }
 ! { dg-require-effective-target vect_double }
 ! { dg-require-effective-target sse2 }
-! { dg-options "-O3 -ffast-math -msse2 -fpredictive-commoning -ftree-vectorize
-fdump-tree-optimized" }
+! { dg-options "-mtune=generic -O3 -ffast-math -msse2 -fpredictive-commoning
-ftree-vectorize -fdump-tree-optimized" }


 ******* RESID COMPUTES THE RESIDUAL:  R = V - AU

[Bug testsuite/59494] [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-12-13
     Ever confirmed|0                           |1

--- Comment #1 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Forgot to add the pointer to
http://gcc.gnu.org/ml/gcc-regression/2013-10/msg00205.html.

[Bug testsuite/59494] [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.9.0

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Well, I don't think configuring with arbitrary --with-cpu/tune is a supported
configuration for zero testsuite FAILs ...?

[Bug testsuite/59494] [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
True, I guess adding -mtune=generic doesn't hurt though, it will be still
broken with --target_board=unix/-mtune=core2 or similar testing, but at least
it will FAIL less often.  For core* tuning the difference is e.g. in reassoc
(different reassoc widths, fairly recent change), and starting with probably
r160636 different IVopts decisions based on different costs.  I believe that is
the only reason for the different number of vector additions.

[Bug testsuite/59494] [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Alternatively we should just remove that scan-tree-dump-times from the test
altogether, counting additions where either lhs or rhs1 of addition has vect_
in name is quite flawed, we have lots of other vector additions that just use
anonymous SSA_NAMEs without any name (_114 etc.), but still have vector types.

Counting addps etc. insns doesn't work though, there are 23 before and after
the r156043 change.

Richard, what is your preference here?

[Bug testsuite/59494] [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 31820
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31820&action=edit
gcc49-pr59494.patch

Fix that tries to capture the r156042 to r156043 changes in the *.pcom dump
details.

[Bug testsuite/59494] [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

--- Comment #8 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
With the patch in comment 7 the test succeeds for

Target: x86_64-apple-darwin10.8.0
Configured with: ../work/configure --prefix=/opt/gcc/gcc4.9w
--enable-languages=c,c++,fortran,objc,obj-c++,ada,java,lto --with-gmp=/opt/mp
--with-system-zlib --with-isl=/opt/mp --enable-lto --enable-plugin

and

Target: x86_64-apple-darwin13.0.0
Configured with: ../work/configure --prefix=/opt/gcc/gcc4.9w
--enable-languages=c,c++,fortran,objc,obj-c++,ada,java,lto --with-gmp=/opt/mp
--with-system-zlib --with-isl=/opt/mp --enable-lto --enable-plugin
--with-arch=corei7 --with-cpu=corei7

(tested with
make -k check-gfortran RUNTESTFLAGS="vect.exp=fast-math-mgrid-resid.f
--target_board=unix'{-m32,-m64}'").

[Bug testsuite/59494] [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Author: jakub
Date: Tue Jan 14 09:00:30 2014
New Revision: 206598

URL: http://gcc.gnu.org/viewcvs?rev=206598&root=gcc&view=rev
Log:
    PR testsuite/59494
    * gfortran.dg/vect/fast-math-mgrid-resid.f: Change
    -fdump-tree-optimized to -fdump-tree-pcom-details in dg-options and
    cleanup-tree-dump from optimized to pcom.  Remove scan-tree-dump-times
    for vect_\[^\\n\]*\\+, add scan-tree-dump-times for no suitable chains and
    Executing predictive commoning without unrolling.

Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gfortran.dg/vect/fast-math-mgrid-resid.f

[Bug testsuite/59494] [4.9 Regression] FAIL: gfortran.dg/vect/fast-math-mgrid-resid.f scan-tree-dump-times optimized "vect_[^\\n]*\\+ " 13

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59494

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

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

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.