topic C vs Fortran documentation in Intel® oneAPI Math Kernel Library

C vs Fortran documentation

Weeks__John — Thu, 15 Oct 2020 23:57:41 GMT

Using MKL in C 2020 update 2

In looking for C documentation for the ?ggevx family of eigenvalue routines, I found that my headers didn't match the documentation. I finally figured out that my headers are C headers for the Fortran routines; if I look at the MKL C documentation, I find LAPACKE_?ggevx() with quite different calling signature.

A quick scan of the C documentation didn't give me an overview of this LAPACKE set of functions. Where would I find it? Is this the preferred way to call from C/C++?

Thanks. I feel like I'm going to be embarrassed when I get the answer, but I'm too old to care

Re: C vs Fortran documentation

mecej4 — Fri, 16 Oct 2020 11:57:39 GMT

There are three ways to call MKL Lapack routines from C:

Call the Fortran callable routine, keeping in mind that your C call should pass arguments in the same way as Fortran does;
The LapackE C interface with work arrays allocated, used and de-allocated inside MKL, and
The LapackE C interface with work array related arguments in the argument list.

For details, see the MKL documentation .

Re:C vs Fortran documentation

Gennady_F_Intel — Fri, 16 Oct 2020 15:52:21 GMT

and you may find out many of such examples into mklroot\examples directory

Re: C vs Fortran documentation

Weeks__John — Thu, 29 Oct 2020 17:18:44 GMT

Thank you to you both, and apologies for the late reply.