topic MKL v.2020 has been released in Intel® oneAPI Math Kernel Library

Signature of v?Fdim, v?Fmax, ...

ingu — Fri, 14 Dec 2018 14:03:03 GMT

Hello,

why is the second array of the functions v?Fdim, v?Fmax, v?Fmin etc. not constant?

The signature in the file mkl_vml_functions.h is:

_Mkl_Api(void, vdFdim, (const MKL_INT n, const double a[], double r1[], double r2[]))
_Mkl_Api(void, vdFmax, (const MKL_INT n, const double a[], double r1[], double r2[]))
_Mkl_Api(void, vdFmin, (const MKL_INT n, const double a[], double r1[], double r2[]))

I think it would be correct as:

_Mkl_Api(void, vdFdim, (const MKL_INT n, const double a[], const double b[], double r[]))
_Mkl_Api(void, vdFmax, (const MKL_INT n, const double a[], const double b[], double r[]))
_Mkl_Api(void, vdFmin, (const MKL_INT n, const double a[], const double b[], double r[]))

Kind regards

Yyes, you are right. the

Gennady_F_Intel — Sat, 15 Dec 2018 05:07:00 GMT

Yes, you are right. the issue has been already raised and escalated. the fix will be available into on the next updates.

Tank you! Then I'm looking

ingu — Mon, 17 Dec 2018 07:48:48 GMT

Tank you! Then I'm looking forward to the next update and I will use until then a const cast.

yes, we will update this

Gennady_F_Intel — Mon, 17 Dec 2018 09:11:38 GMT

yes, we will update this thread when the fix of the problem will be available.

MKL v.2020 has been released

Gennady_F_Intel — Tue, 17 Dec 2019 03:17:42 GMT

MKL v.2020 has been released and contains the fix of this problem either.