[Bug libfortran/26985] New: incorrect matmul result

[Bug libfortran/26985] New: incorrect matmul result

[fxcoudert at gcc dot gnu dot org]

$ cat fail.f90 
  implicit none
  real(8) :: a(2,9), b(9,7), c(2,7)
  integer :: i, j

  a = 0.d0
  b = 0.d0
  c = 1789789.d0
  c(:,1:7:2) = matmul(a,b(:,1:7:2))
  do i = 1, 7
    print *, c(:,i)
  end do
  end
$ ifort fail.f90 && ./a.out   
  0.000000000000000E+000  0.000000000000000E+000
   1789789.00000000        1789789.00000000     
  0.000000000000000E+000  0.000000000000000E+000
   1789789.00000000        1789789.00000000     
  0.000000000000000E+000  0.000000000000000E+000
   1789789.00000000        1789789.00000000     
  0.000000000000000E+000  0.000000000000000E+000
$ gfortran fail.f90 && ./a.out
   0.00000000000000        0.00000000000000     
   0.00000000000000        0.00000000000000     
   0.00000000000000        0.00000000000000     
   0.00000000000000        0.00000000000000     
   1789789.00000000        1789789.00000000     
   1789789.00000000        1789789.00000000     
   1789789.00000000        1789789.00000000


-- 
           Summary: incorrect matmul result
           Product: gcc
           Version: 4.2.0
            Status: UNCONFIRMED
          Severity: major
          Priority: P3
         Component: libfortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: fxcoudert at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] incorrect matmul result

[fxcoudert at gcc dot gnu dot org]

------- Comment #1 from fxcoudert at gcc dot gnu dot org  2006-04-02 10:49 -------
Analysis and patch idea here:
http://gcc.gnu.org/ml/fortran/2006-04/msg00029.html


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |fxcoudert at gcc dot gnu dot
                   |dot org                     |org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2006-04-02 10:49:37
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] incorrect matmul result

[tkoenig at gcc dot gnu dot org]

------- Comment #2 from tkoenig at gcc dot gnu dot org  2006-04-02 18:41 -------
The test case also exposes a problem with the way that
the result is stored.

Look at this:
$ cat matmul.f90
program main
  implicit none
  real(8) :: a(2,9), b(9,7), c(2,7)
  integer :: i, j

  a = 1.d0
  b = 2.d0
  c = 1789789.d0
  c(:,1:7:2) = matmul(a,b(:,1:7:2))
  do i = 1, 7
     print *, c(:,i)
  end do
end program main

$ gfortran matmul.f90 && ./a.out
   18.0000000000000        18.0000000000000
   0.00000000000000        0.00000000000000
   18.0000000000000        18.0000000000000
   0.00000000000000        0.00000000000000
   1789807.00000000        1789807.00000000
   1789789.00000000        1789789.00000000
   1789807.00000000        1789807.00000000
$ ifort matmul.f90 && ./a.out
   18.0000000000000        18.0000000000000
   1789789.00000000        1789789.00000000
   18.0000000000000        18.0000000000000
   1789789.00000000        1789789.00000000
   18.0000000000000        18.0000000000000
   1789789.00000000        1789789.00000000
   18.0000000000000        18.0000000000000


-- 

tkoenig at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] incorrect matmul result

[tkoenig at gcc dot gnu dot org]

------- Comment #3 from tkoenig at gcc dot gnu dot org  2006-04-02 19:21 -------
(In reply to comment #2)
> The test case also exposes a problem with the way that
> the result is stored.

Actually, it doesn't.  Storing the result is OK, clearing beforehand isn't.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] incorrect matmul result

[tkoenig at gcc dot gnu dot org]

------- Comment #4 from tkoenig at gcc dot gnu dot org  2006-04-29 16:30 -------
Hi FX,

would you care to replace rystride == ycount with rystride == xcount
as a fix for this, independent of the matmul BLAS issue?

I think this is definitely worth it, especially for 4.1.

Pre-approved if you do so :-)

   Thomas


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] incorrect matmul result

[fxcoudert at gcc dot gnu dot org]

------- Comment #6 from fxcoudert at gcc dot gnu dot org  2006-05-05 11:08 -------
Subject: Bug 26985

Author: fxcoudert
Date: Fri May  5 11:08:23 2006
New Revision: 113552

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=113552
Log:
        PR libfortran/26985

        * m4/matmul.m4: Correct the condition for the memset call,
        and remove the unneeded call to size0.
        * generated/matmul_r4.c: Regenerate.
        * generated/matmul_r8.c: Regenerate.
        * generated/matmul_r10.c: Regenerate.
        * generated/matmul_r16.c: Regenerate.
        * generated/matmul_c4.c: Regenerate.
        * generated/matmul_c8.c: Regenerate.
        * generated/matmul_c10.c: Regenerate.
        * generated/matmul_c16.c: Regenerate.
        * generated/matmul_i4.c: Regenerate.
        * generated/matmul_i8.c: Regenerate.
        * generated/matmul_i16.c: Regenerate.

        * gfortran.dg/matmul_2.f90: New test.

Added:
    trunk/gcc/testsuite/gfortran.dg/matmul_2.f90
Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/generated/matmul_c10.c
    trunk/libgfortran/generated/matmul_c16.c
    trunk/libgfortran/generated/matmul_c4.c
    trunk/libgfortran/generated/matmul_c8.c
    trunk/libgfortran/generated/matmul_i16.c
    trunk/libgfortran/generated/matmul_i4.c
    trunk/libgfortran/generated/matmul_i8.c
    trunk/libgfortran/generated/matmul_r10.c
    trunk/libgfortran/generated/matmul_r16.c
    trunk/libgfortran/generated/matmul_r4.c
    trunk/libgfortran/generated/matmul_r8.c
    trunk/libgfortran/m4/matmul.m4


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] incorrect matmul result

[fxcoudert at gcc dot gnu dot org]

------- Comment #5 from fxcoudert at gcc dot gnu dot org  2006-05-05 11:07 -------
(In reply to comment #4)
> would you care to replace rystride == ycount with rystride == xcount
> as a fix for this, independent of the matmul BLAS issue?
> 
> I think this is definitely worth it, especially for 4.1.
> 
> Pre-approved if you do so :-)

Commiting on mainline first, after regtesting on i686-linux and adding the
corresponding testcase. I will commit on 4.1 branch in a few days.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] [4.1 only] incorrect matmul result

[fxcoudert at gcc dot gnu dot org]

-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch
      Known to fail|                            |4.1.1
      Known to work|                            |4.2.0
            Summary|incorrect matmul result     |[4.1 only] incorrect matmul
                   |                            |result
   Target Milestone|---                         |4.1.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] [4.1 only] incorrect matmul result

[fxcoudert at gcc dot gnu dot org]

------- Comment #7 from fxcoudert at gcc dot gnu dot org  2006-05-07 13:02 -------
Subject: Bug 26985

Author: fxcoudert
Date: Sun May  7 13:02:39 2006
New Revision: 113600

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=113600
Log:
        PR libfortran/26985

        * m4/matmul.m4: Correct the condition for the memset call,
        and remove the unneeded call to size0.
        * generated/matmul_r4.c: Regenerate.
        * generated/matmul_r8.c: Regenerate.
        * generated/matmul_r10.c: Regenerate.
        * generated/matmul_r16.c: Regenerate.
        * generated/matmul_c4.c: Regenerate.
        * generated/matmul_c8.c: Regenerate.
        * generated/matmul_c10.c: Regenerate.
        * generated/matmul_c16.c: Regenerate.
        * generated/matmul_i4.c: Regenerate.
        * generated/matmul_i8.c: Regenerate.
        * generated/matmul_i16.c: Regenerate.

        * gfortran.dg/matmul_2.f90: New test.

Added:
    branches/gcc-4_1-branch/gcc/testsuite/gfortran.dg/matmul_2.f90
      - copied unchanged from r113552,
trunk/gcc/testsuite/gfortran.dg/matmul_2.f90
Modified:
    branches/gcc-4_1-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_1-branch/libgfortran/ChangeLog
    branches/gcc-4_1-branch/libgfortran/generated/matmul_c10.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_c16.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_c4.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_c8.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_i16.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_i4.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_i8.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_r10.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_r16.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_r4.c
    branches/gcc-4_1-branch/libgfortran/generated/matmul_r8.c
    branches/gcc-4_1-branch/libgfortran/m4/matmul.m4


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985

[Bug libfortran/26985] incorrect matmul result

[fxcoudert at gcc dot gnu dot org]

------- Comment #8 from fxcoudert at gcc dot gnu dot org  2006-05-07 15:49 -------
Fixed on both 4.1 and mainline.


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
           Keywords|patch                       |
      Known to fail|4.1.1                       |
      Known to work|4.2.0                       |4.2.0 4.1.1
         Resolution|                            |FIXED
            Summary|[4.1 only] incorrect matmul |incorrect matmul result
                   |result                      |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26985