[Bug middle-end/42180] New: compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[Bug middle-end/42180] New: compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[howarth at nitro dot med dot uc dot edu]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 5044 bytes --]

The polyhedron 2005 benchmark induct.f90 ICEs the gfortran compiler in current
gcc trunk when compiled with -ffast-math -O2 -fgraphite-identity...

gfortran -ffast-math -O2 -fgraphite-identity induct.f90 -o induct
induct.f90: In function ‘__mcc_m_MOD_mutual_ind_cir_cir_coils’:
induct.f90:777:0: internal compiler error: Segmentation fault

A backtrace shows...

(gdb) r induct.f90 -fPIC -quiet -dumpbase induct.f90
-mmacosx-version-min=10.6.2 -mtune=generic -auxbase induct -O2 -version
-ffast-math -fgraphite-identity -fintrinsic-modules-path
/sw/lib/gcc4.5/lib/gcc/x86_64-apple-darwin10.2.0/4.5.0/finclude -o
/var/folders/1C/1CdoNxmNFHyOIjNBLNuJh++++TM/-Tmp-//cczMDWLV.s
Starting program:
/sw/lib/gcc4.5/libexec/gcc/x86_64-apple-darwin10.2.0/4.5.0/f951 induct.f90
-fPIC -quiet -dumpbase induct.f90 -mmacosx-version-min=10.6.2 -mtune=generic
-auxbase induct -O2 -version -ffast-math -fgraphite-identity
-fintrinsic-modules-path
/sw/lib/gcc4.5/lib/gcc/x86_64-apple-darwin10.2.0/4.5.0/finclude -o
/var/folders/1C/1CdoNxmNFHyOIjNBLNuJh++++TM/-Tmp-//cczMDWLV.s
Reading symbols for shared libraries .++++++++++++..... done
GNU Fortran (GCC) version 4.5.0 20091125 (experimental)
(x86_64-apple-darwin10.2.0)
        compiled by GNU C version 4.5.0 20091125 (experimental), GMP version
4.3.1, MPFR version 2.4.1, MPC version 0.8
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
GNU Fortran (GCC) version 4.5.0 20091125 (experimental)
(x86_64-apple-darwin10.2.0)
        compiled by GNU C version 4.5.0 20091125 (experimental), GMP version
4.3.1, MPFR version 2.4.1, MPC version 0.8
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096

Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x0000000000000000
0x000000010053b1e2 in rewrite_commutative_reductions_out_of_ssa_loop
(loop=<value temporarily unavailable, due to optimizations>, reductions=<value
temporarily unavailable, due to optimizations>) at
../../gcc-4.5-20091125/gcc/graphite-sese-to-poly.c:2474
2474      e = split_block (bb, gsi_stmt (gsi));
(gdb) bt
#0  0x000000010053b1e2 in rewrite_commutative_reductions_out_of_ssa_loop
(loop=<value temporarily unavailable, due to optimizations>, reductions=<value
temporarily unavailable, due to optimizations>) at
../../gcc-4.5-20091125/gcc/graphite-sese-to-poly.c:2474
#1  0x0000000100543205 in build_poly_scop (scop=<value temporarily unavailable,
due to optimizations>) at
../../gcc-4.5-20091125/gcc/graphite-sese-to-poly.c:2819
#2  0x000000010051ec3f in graphite_transform_loops () at
../../gcc-4.5-20091125/gcc/graphite.c:273
#3  0x00000001007c185a in graphite_transforms () at
../../gcc-4.5-20091125/gcc/tree-ssa-loop.c:300
#4  0x0000000100601681 in execute_one_pass (pass=0x100c68660) at
../../gcc-4.5-20091125/gcc/passes.c:1522
#5  0x000000010060197d in execute_pass_list (pass=0x100c68660) at
../../gcc-4.5-20091125/gcc/passes.c:1577
#6  0x000000010060198f in execute_pass_list (pass=0x100c683c0) at
../../gcc-4.5-20091125/gcc/passes.c:1578
#7  0x000000010060198f in execute_pass_list (pass=0x100c67a00) at
../../gcc-4.5-20091125/gcc/passes.c:1578
#8  0x00000001007332a4 in tree_rest_of_compilation (fndecl=0x141f0fd00) at
../../gcc-4.5-20091125/gcc/tree-optimize.c:411
#9  0x00000001008eb9cb in cgraph_expand_function (node=0x141d0aee0) at
../../gcc-4.5-20091125/gcc/cgraphunit.c:1178
#10 0x00000001008edaca in cgraph_optimize () at
../../gcc-4.5-20091125/gcc/cgraphunit.c:1245
#11 0x00000001008ee0df in cgraph_finalize_compilation_unit () at
../../gcc-4.5-20091125/gcc/cgraphunit.c:1107
#12 0x0000000100595a26 in write_global_declarations () at
../../gcc-4.5-20091125/gcc/langhooks.c:309
#13 0x00000001006c8772 in toplev_main (argc=18, argv=0x7fff5fbfd8b8) at
../../gcc-4.5-20091125/gcc/toplev.c:1061
#14 0x0000000100001854 in start ()

Using built-in specs.
COLLECT_GCC=gfortran
COLLECT_LTO_WRAPPER=/sw/lib/gcc4.5/libexec/gcc/x86_64-apple-darwin10.2.0/4.5.0/lto-wrapper
Target: x86_64-apple-darwin10.2.0
Configured with: ../gcc-4.5-20091125/configure --prefix=/sw
--prefix=/sw/lib/gcc4.5 --mandir=/sw/share/man --infodir=/sw/share/info
--enable-languages=c,c++,fortran,objc,obj-c++,java --with-gmp=/sw
--with-libiconv-prefix=/sw --with-ppl=/sw --with-cloog=/sw --with-mpc=/sw
--with-system-zlib --x-includes=/usr/X11R6/include --x-libraries=/usr/X11R6/lib
--disable-libjava-multilib
Thread model: posix
gcc version 4.5.0 20091125 (experimental) (GCC)


-- 
           Summary: compiling induct.f90 with  -ffast-math -O2 -fgraphite-
                    identity ICEs gfortran
           Product: gcc
           Version: 4.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: howarth at nitro dot med dot uc dot edu
 GCC build triplet: x86_64-apple-darwin10
  GCC host triplet: x86_64-apple-darwin10
GCC target triplet: x86_64-apple-darwin10


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180

[Bug middle-end/42180] compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[spop at gcc dot gnu dot org]

------- Comment #1 from spop at gcc dot gnu dot org  2009-12-18 06:33 -------
Mine.


-- 

spop at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |spop at gcc dot gnu dot org
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2009-12-18 06:33:17
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180

[Bug middle-end/42180] compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[spop at gcc dot gnu dot org]

------- Comment #2 from spop at gcc dot gnu dot org  2009-12-18 07:43 -------
Reduced testcase:

module mcc_m
integer, parameter, private :: longreal = selected_real_kind(15,90)
contains
      subroutine mutual_ind_cir_cir_coils (r1, r2, x12, y12, z12, l1, l2,
turns1, turns2,  &
                                            mu, rotate, m, l12)
      real (kind = longreal), intent(out) :: l12
      real (kind = longreal), dimension(1:9), save :: z1g, z2g, zw
gauss:do i = 1, 9
          theta_l12 = 0.0_longreal
theta1:   do n1 = 1, 2*m
              theta_1 = pi*real(n1,longreal)/real(m,longreal)
theta2:       do n2 = 1, 2*m
                  numerator = -sin(theta_1)*tvx + cos(theta_1)*tvy
                  theta_l12 = theta_l12 + numerator/denominator
              end do theta2
           end do theta1
           l12 = l12 + zw(i)*theta_l12
      end do gauss
      l12 = coefficient * l12
      end subroutine mutual_ind_cir_cir_coils
end module mcc_m


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180

[Bug middle-end/42180] compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[spop at gcc dot gnu dot org]

------- Comment #3 from spop at gcc dot gnu dot org  2009-12-18 07:55 -------
Subject: Bug 42180

Author: spop
Date: Fri Dec 18 07:55:05 2009
New Revision: 155338

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=155338
Log:
Fix PR42180.

2009-12-18  Sebastian Pop  <sebastian.pop@amd.com>

        PR middle-end/42180
        * graphite-sese-to-poly.c (split_reduction_stmt): Do not split blocks
        that contain only one statement.

Added:
    branches/graphite/gcc/testsuite/gfortran.dg/graphite/pr42180.f90
Modified:
    branches/graphite/gcc/ChangeLog.graphite
    branches/graphite/gcc/graphite-sese-to-poly.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180

[Bug middle-end/42180] compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[spop at gcc dot gnu dot org]

------- Comment #4 from spop at gcc dot gnu dot org  2009-12-18 07:56 -------
Still not fully fixed, I'm reducing another testcase that fails a bit later in
induct.f90


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180

[Bug middle-end/42180] compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[spop at gcc dot gnu dot org]

------- Comment #6 from spop at gcc dot gnu dot org  2009-12-18 08:29 -------
Fixed.


-- 

spop at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180

[Bug middle-end/42180] compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[spop at gcc dot gnu dot org]

------- Comment #5 from spop at gcc dot gnu dot org  2009-12-18 08:29 -------
Subject: Bug 42180

Author: spop
Date: Fri Dec 18 08:28:45 2009
New Revision: 155339

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=155339
Log:
Fix PR42180.

2009-12-18  Sebastian Pop  <sebastian.pop@amd.com>

        PR middle-end/42180
        * graphite-sese-to-poly.c (follow_ssa_with_commutative_ops): Handle
        GIMPLE_CALL.

        * testsuite/gfortran.dg/graphite/pr42180.f90: Add compile flags.

Modified:
    branches/graphite/gcc/ChangeLog.graphite
    branches/graphite/gcc/graphite-sese-to-poly.c
    branches/graphite/gcc/testsuite/gfortran.dg/graphite/pr42180.f90


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180

[Bug middle-end/42180] compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[spop at gcc dot gnu dot org]

------- Comment #8 from spop at gcc dot gnu dot org  2009-12-23 07:51 -------
Subject: Bug 42180

Author: spop
Date: Wed Dec 23 07:51:05 2009
New Revision: 155423

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=155423
Log:
Fix PR42180.

2009-12-18  Sebastian Pop  <sebastian.pop@amd.com>

        PR middle-end/42180
        * graphite-sese-to-poly.c (follow_ssa_with_commutative_ops): Handle
        GIMPLE_CALL.

        * testsuite/gfortran.dg/graphite/pr42180.f90: Add compile flags.

Modified:
    trunk/gcc/ChangeLog.graphite
    trunk/gcc/graphite-sese-to-poly.c
    trunk/gcc/testsuite/gfortran.dg/graphite/pr42180.f90


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180

[Bug middle-end/42180] compiling induct.f90 with -ffast-math -O2 -fgraphite-identity ICEs gfortran

[spop at gcc dot gnu dot org]

------- Comment #7 from spop at gcc dot gnu dot org  2009-12-23 07:51 -------
Subject: Bug 42180

Author: spop
Date: Wed Dec 23 07:50:53 2009
New Revision: 155422

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=155422
Log:
Fix PR42180.

2009-12-18  Sebastian Pop  <sebastian.pop@amd.com>

        PR middle-end/42180
        * graphite-sese-to-poly.c (split_reduction_stmt): Do not split blocks
        that contain only one statement.

Added:
    trunk/gcc/testsuite/gfortran.dg/graphite/pr42180.f90
Modified:
    trunk/gcc/ChangeLog.graphite
    trunk/gcc/graphite-sese-to-poly.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42180