Intel® oneAPI Math Kernel Library
Ask questions and share information with other developers who use Intel® Math Kernel Library.
Announcements
The Intel sign-in experience has changed to support enhanced security controls. If you sign in, click here for more information.
6743 Discussions

different BLAS and LAPACK functions between OpenMP and Intel TBB

semihfk
Beginner
333 Views

Hello everyone,

I generate a C code from my simulink model. That works fine. I also call the BLAS and Lapack libraries from Intel oneAPI MKL in simulink model in order to speed up the matrix operations as much as possible. But I noticed some BLAS functions are not used in C code. They are level 1 and 2 BLAS functions. The level 3 BLAS functions are successfully used in C code.

My simulink model includes some matrix-vector multiplications, vector additions and scalar-vector multiplications as well which require BLAS functions to be used for level 1 and level 2. The BLAS functions cblas_dgemm and cblas_dtrsm as well as the Lapack function lapacke_dgetrf work well.

About this issue I read in the Intel MKL Developer Guide that the usable functions are different in threading with OpenMP and Intel TBB. There are no the Level 1 and Level 2 functions threaded with Intel TBB. I got it from that that I should use OpenMP in order to use these functions as well.

My question is whether the compiler used in matlab has any influence on it. If so, which compilers should I use? I have been using the microsoft visual c++ 2019 compiler in matlab.

What should I do to use BLAS level 1 and 2 functions as well?

 

I would be grateful for any advice. Thanks.
0 Kudos
1 Reply
NoorjahanSk_Intel
Moderator
305 Views

Hi,

Thanks for reaching out to us.


Since this is a duplicate thread of

https://community.intel.com/t5/Intel-oneAPI-Threading-Building/Intel-oneAPI-MKL-BLAS-and-LAPACK-functions/m-p/1346041/highlight/true

We will no longer monitor this thread. We will continue addressing

this issue in the other thread


Thanks & Regards,

Noorjahan.


Reply