Hi Sturla, Quoting Sturla Molden <stu...@molden.no>:
> Den 10.03.2012 22:56, skrev Sturla Molden: >> >> I am not sure why NumPy uses f2c'd routines instead of a dependency >> on BLAS and LAPACK like SciPy. > > Actually, np.dot does depend on the CBLAS interface to BLAS (_dotblas.c). > > But the lapack methods in lapack_lite seems to use f2c'd code. I am > not sure if they will use an optimized BLAS or just link to f2c's > BLAS in blas_lite.c. > > If the intention is to avoid Fortran dependency in NumPy, I am not > sure why this is better than a dependency on CBLAS and LAPACKE. > Thanks for the more complete information. Now I understand better why it is more difficult to access the underlying libraries when using numpy instead of scipy. My main objective was to avoid having to ship libraries with my python extension modules. Using those already available via numpy seemed to me the most natural option since my C extension depends on numpy but not on scipy. In addition, people using my extension module could benefit from better libraries than those I would be able to ship. Eliminating the fortran dependency appeared to me as an added bonus: I have a 64 bit intel fortran compiler license for the only reason of not being limited by a fortran dependency. Best regards, Armando _______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion