[Bug tree-optimization/18168] New: SPEC CPU2000 173.applu tree-loop-linear failure

[Bug tree-optimization/18168] New: SPEC CPU2000 173.applu tree-loop-linear failure

[dje at gcc dot gnu dot org]

gcc -O3 -ftree-loop-linear applu.f produces error:
Dependence relation cannot be represented by distance vector.

-- 
           Summary: SPEC CPU2000 173.applu tree-loop-linear failure
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: dje at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: *-*-*
  GCC host triplet: *-*-*
GCC target triplet: *-*-*


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dje at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|SPEC CPU2000 173.applu tree-|SPEC CPU2000 173.applu tree-
                   |loop-linear failure         |loop-linear ICE


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dje at gcc dot gnu dot org]

------- Additional Comments From dje at gcc dot gnu dot org  2004-10-28 15:31 -------
gfortran -c -o applu.o -O3 -mcpu=power4 -ftree-loop-linear applu.f
applu.f: In function 'buts':
applu.f:633: internal compiler error: in build_classic_dist_vector, at tree-
data-ref.c:1848

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dberlin at dberlin dot org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-10-28 15:31:35
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-28 15:52 -------
Isn't this related to PR 17672.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dberlin at dberlin dot org]

------- Additional Comments From dberlin at dberlin dot org  2004-10-28 16:31 -------
Subject: Re:  SPEC CPU2000 173.applu tree-loop-linear
 ICE

This is the minimal testcase for the problem that i came up with
       subroutine buts ( ldmx, ldmy, ldmz,
      $                  v,
      $                  d, udx, udy, udz )
       dimension v( 5, ldmx, ldmy, *),
      $          udz( 5, 5, ldmx, ldmy, * )
       dimension tmat(5,5), tv(5)
       do k = nz-1, 2, -1
          do j = ny-1, 2, -1
             do i = nx-1, 2, -1
                do ip = 1, 4
                   do m = ip+1, 5
                      do l = ip+1, 5
                      end do
                      tv( m ) = tv( m ) -
      $                  tv( ip )  *
      $                 tmp
                   end do
                end do
                do m = 5, 1, -1
                   do l = m+1, 5
                   end do
                   v( m, i, j, k ) = v( m, i, j, k )
      $            - tv( m )
                end do
             end do
          end do
       end do
       end



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dberlin at dberlin dot org]

------- Additional Comments From dberlin at dberlin dot org  2004-10-28 16:36 -------
Subject: Re:  SPEC CPU2000 173.applu tree-loop-linear
 ICE

No, not at all actually.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-10-28 16:51 -------
Reduced testcase:

================================
      subroutine buts ( v )
      dimension v( 5)
      do i = 1, 5
         do j = 1, 5
            do k = 1, 5
              v( i ) = v( 1 )
            end do
         end do
      end do
      end
================================

To me this looks very much like comment #4 from PR 17672.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |reichelt at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dberlin at dberlin dot org]

------- Additional Comments From dberlin at dberlin dot org  2004-10-28 16:55 -------
Subject: Re:  SPEC CPU2000 173.applu tree-loop-linear
 ICE

> ------- Additional Comments From reichelt at gcc dot gnu dot org  2004-10-28 16:51 -------
> Reduced testcase:
>
> ================================
>      subroutine buts ( v )
>      dimension v( 5)
>      do i = 1, 5
>         do j = 1, 5
>            do k = 1, 5
>              v( i ) = v( 1 )
>            end do
>         end do
>      end do
>      end
> ================================
>
> To me this looks very much like comment #4 from PR 17672.

Looks can be deceiving :)
Try the patch i posted with the blob of code from comment 4, and you'll 
see that blob passes with the patch, but fails with the above.




-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dberlin at gcc dot gnu dot org]

------- Additional Comments From dberlin at gcc dot gnu dot org  2004-10-30 02:03 -------
The problem here is that find_data_references_in_loop is handing us data
references that involve variables only outside the loop.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dberlin at gcc dot gnu dot org]

------- Additional Comments From dberlin at gcc dot gnu dot org  2004-10-30 03:10 -------
I justed fixed this, so i might as well assign it to myself

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |dberlin at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-10-30 13:06 -------
I assume this is a regression?

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dberlin at dberlin dot org]

------- Additional Comments From dberlin at dberlin dot org  2004-10-30 13:15 -------
Subject: Re:  SPEC CPU2000 173.applu tree-loop-linear
 ICE

It only pops up with -ftree-loop-linear, which is a new option for 4.0.
So no, it's probably not a regression.

On Sat, 30 Oct 2004, giovannibajo at libero dot it wrote:

>
> ------- Additional Comments From giovannibajo at libero dot it  2004-10-30 13:06 -------
> I assume this is a regression?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-11-01 18:08 -------
Subject: Bug 18168

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	dberlin@gcc.gnu.org	2004-11-01 18:08:03

Modified files:
	gcc            : ChangeLog lambda-code.c tree-data-ref.c 
	                 tree-loop-linear.c tree-optimize.c 
Added files:
	gcc/testsuite/gcc.dg/tree-ssa: ltrans-1.c ltrans-2.c ltrans-3.c 
	                               ltrans-4.c ltrans-5.c 

Log message:
	2004-10-16  Daniel Berlin  <dberlin@dberlin.org>
	
	Fix PR tree-optimization/17672
	Fix PR tree-optimization/18168
	
	* lambda-code.c (lambda_lattice_compute_base): Fix reversed
	assert test.
	(gcc_tree_to_linear_expression): Add extra to existing constant.
	(depth_of_nest): Factor out function used in various places.
	(gcc_loop_to_lambda_loop): Clean up code a little bit. No
	functional changes.
	(find_induction_var_from_exit_cond): Stop guessing, and just
	get the right answer :).
	(gcc_loopnest_to_lambda_loopnest): Remove useless pre-allocation.
	Print out message about result of attempt to create perfect nest.
	(lbv_to_gcc_expression): Add type argument, use it to do math
	and induction variable creation.
	(lle_to_gcc_expression): Ditto.
	(lambda_loopnest_to_gcc_loopnest): Create new iv with same type as
	oldiv. Pass type argument to lle_to_gcc_expression and
	lbv_to_gcc_expression.
	Reset number of iterations after transformation.
	(perfect_nestify): Remove useless pre-allocation, and cleanup
	a small amount.
	
	* tree-data-ref.c (build_classic_dist_vector): Return false for
	dependences completely outside of the loop nest we asked about.
	(build_classic_dir_vector): Ditto.
	(compute_data_dependences_for_loop): Only add dependence relations
	inside the loop we asked about.
	
	* tree-loop-linear.c (linear_transform_loops): Use DDR_SIZE_VECT.
	Compute immediate uses.
	
	* tree-optimize.c: Move linear_transform_loops to before ivcanon.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.6120&r2=2.6121
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/lambda-code.c.diff?cvsroot=gcc&r1=2.16&r2=2.17
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-data-ref.c.diff?cvsroot=gcc&r1=2.14&r2=2.15
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-loop-linear.c.diff?cvsroot=gcc&r1=2.2&r2=2.3
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-optimize.c.diff?cvsroot=gcc&r1=2.59&r2=2.60
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/tree-ssa/ltrans-1.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/tree-ssa/ltrans-2.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/tree-ssa/ltrans-3.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/tree-ssa/ltrans-4.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/tree-ssa/ltrans-5.c.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[dberlin at gcc dot gnu dot org]

------- Additional Comments From dberlin at gcc dot gnu dot org  2004-11-01 19:22 -------
Fixed

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168

[Bug tree-optimization/18168] SPEC CPU2000 173.applu tree-loop-linear ICE

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18168