Intel® oneAPI Math Kernel Library
Ask questions and share information with other developers who use Intel® Math Kernel Library.

MKL + gfortran on ARM processors (e.g. Raspberry Pi 3, quad core ARMv8 64bit)

moortgatgmail_com
3,453 Views

The short question: is it possible to use MKL on current ARM processors?

The slightly longer version: recent ARM processors, such as the one in the subject line, are quite fast and support full Linux installations (e.g. Raspbian, Ubuntu and others), as well as gcc, gfortran, and blas/lapack libraries. Perhaps it's not surprising that Intel "Compiler+MKL" Composer Suites are not (fully) supported on ARM architectures. But the combination of stand-alone MKL (as here https://software.intel.com/en-us/articles/intel-mkl-111-install-guide ), linked against gfortran seems like it should be possible. Is it, though? Any suggestions or experiences by others would be appreciated.

 

 

 

 

0 Kudos
2 Replies
Gennady_F_Intel
Moderator
3,453 Views

The answer is that not possible. Here is from MKL's system requirements : " The Intel MKL supports the IA-32 and Intel® 64 architectures. For a complete explanation of these architecture names please read the following article:Intel Architecture Platform Terminology for Development Tools".  

 

0 Kudos
moortgatgmail_com
3,453 Views

I read those system requirements, and realize that MKL on ARM may not be "supported", but I'm not sure that means it is not "possible". If the MKL libraries are fortran compliant, it may be possible to compile the source/library with the gcc/gfortran tools available for ARM.  OpenMP, MPICH, Mathematica etc have also all been ported to ARM. I was hoping that perhaps others have tinkered with / tried to make MKL work on this new architecture (with easily more than 2 billion of these processors being used by Android & iPhone users alone). This would open up many interesting applications.

 

 

0 Kudos
Reply