[Octave-bug-tracker] [bug #43908] Some OF packages won't install with 64-bit indexing Octave

[John W. Eaton]

Follow-up Comment #20, bug #43908 (project octave):

For mxe-octave, the --enable-windows-64 uses 64-bit pointers, so the resulting
binary won't work on a 32-bit system.

The --enable-64 option (for mxe-octave) causes the Fortran parts to be built
with 64-bit integers.  This also applies to OpenBLAS, which is not written in
Fortran, but by default uses 32-bit integers to conform to the reference BLAS,
which *is* defined in Fortran.  There is little need for this now unless you
need to do linear algebra on very large arrays.

For Octave itself, we now build with 64-bit indexing by default but configure
detects whether the BLAS uses 32-bit integers.  If so, then we limit the size
of arrays that can be passed to those functions.

As the NEWS file says:


 ** On systems with 64-bit pointers, --enable-64 is now the default and
    Octave always uses 64-bit indexing.  However, if the configure
    script determines that the BLAS library uses 32-bit integers, then
    operations using the following libraries are limited to arrays with
    dimensions that are smaller than 2^31 elements:

      BLAS  LAPACK  QRUPDATE  SuiteSparse  ARPACK

    Additionally, the following libraries use "int" internally, so
    maximum problem sizes are always limited:

      glpk  Qhull



    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?43908>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/