This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: matmul intrinsic procedure
- From: Jan Stanisław Owoc <jsowoc at ucalgary dot ca>
- To: Richard E Maine <Richard dot Maine at nasa dot gov>
- Cc: fortran at gcc dot gnu dot org
- Date: Thu, 30 Jun 2005 15:13:13 -0600 (MDT)
- Subject: Re: matmul intrinsic procedure
I have a test case and the CVS version of GCC 4.1 from June 27th:
[jsowoc@localhost ~]$ gfortran -v
Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ./configure --prefix=/localhost/stow/gcc-4.1-cvs
Thread model: posix
gcc version 4.1.0 20050627 (experimental)
[jsowoc@localhost ~]$ gfortran test.f95 -o test.out
[jsowoc@localhost ~]$ ./test.out
test.out: ../../../libgfortran/generated/matmul_r8.c:170: matmul_r8:
Assertion `a->base == 0' failed.
Aborted
[jsowoc@localhost ~]$ cat test.f95
PROGRAM TEST
IMPLICIT NONE
DOUBLE PRECISION, ALLOCATABLE :: A(:,:),X(:),B(:)
INTEGER, PARAMETER :: n = 3
ALLOCATE ( A(n,n), X(n), B(n))
A = RESHAPE( source = (/ 2.0, 5.0, 3.0, 6.0, 3.0, 7.0, 6.0, 3.0, 2.0 /),
shape = (/3, 3/) )
X = (/ 2.7, 5.4, 8.2 /)
B = MATMUL( A, X )
PRINT *, B
END PROGRAM TEST
[jsowoc@localhost ~]$
Someone got it working with integers on fixed-size arrays, so the problem
is either that I like to allocate arrays once I read in some data (I can't
tell how large a finite-element mesh is up front) and that my force
vectors tend to be floating point numbers rather than integers.
Where/how do I file this as a "bug"?
BTW, the list of intrinsic procedures in the docs does not include
"matmul":
http://gcc.gnu.org/onlinedocs/gcc-4.0.0/gfortran/Intrinsic-Procedures.html
Thanks for all the help.
Jan Owoc