Free Download

Intel® Math Kernel Library (Intel® MKL) for Linux*

Get Your Free Download

Your license includes the full version of the product.

  1. Make sure your system meets the minimum requirements listed on this page.
  2. Complete the registration form.
  3. Download the product.

Register & Download

What’s Included

  • Runtime environment, including:
    • Shared (.so) libraries
  • Software development kit (SDK), including:
    • Supporting utilities
    • Include files and modules
    • Static (.a) libraries
    • Test codes

Minimum System Requirements

Intel MKL supports IA-32 and Intel® 64 architectures.

Hardware Requirements

  • Systems with IA-32 or Intel® 64 architecture or compatible, in particular:
    • Intel® Core™ processor family
    • Intel® Xeon® processor Scalable family
  • 1 GB of RAM per core (2 GB recommended)
  • 1 GB of free hard disk space

Software Requirements

  • Operating systems:
    • Red Hat* Enterprise Linux* 6, 7, 7.5 (IA-32 and Intel® 64 architecture)
    • Red Hat Fedora* core 25, 26 (IA-32 and Intel 64 architecture)
    • SUSE Linux Enterprise Server* 11, 12 (IA-32 and Intel 64 architecture)
    • CentOS* 6.0, 7.1, 7.2
    • Debian* 8, 9 (IA-32 and Intel 64 architecture)
    • Ubuntu* 14.04 LTS, 16.04 LTS, 17.04 LTS, 17.10 LTS (IA-32 and Intel 64 architecture)
    • WindRiver Linux* 7, 8, 9
    • Yocto Project* 1.9, 2.2

Supported C, C++, and Fortran Compilers

  • Intel® Fortran Composer XE for Linux* OS (2016, 2017, 2018)
  • Intel® Fortran Compiler (16.0, 17.0, 18.0)
  • Intel® C++ Composer XE for Linux* OS (2016, 2017, 2018)
  • GNU Compiler Collection 4.9 and later
  • PGI* Compiler version (2015, 2016)

Validated MPI Implementations

  • Intel® MPI Library version 5.1, 2017, 2018 (Intel 64 architecture)
  • MPICH2 version 1.5 (Intel 64 architecture) and 3.1, 3.2
  • Open MPI 1.8.x (Intel 64 architecture)

Complete Specifications
Release Notes

Installation Notes

Before you begin using Intel Math Kernel Library for Linux, complete the following steps.

  1. Install the product in the <parent directory> directory.
  2. Check that the <mkl directory> subdirectory of <parent directory> was created.
  3. To keep multiple versions of the library installed on your system, update your build scripts to point to the correct one.
  4. Check that the following files appear in the <mkl directory>/bin directory:
    mklvars.sh
    mklvars.csh
    Use these files to assign specific Intel MKL values to several environment variables, as explained in Setting Environment Variables.
  5. To understand how the Intel MKL directories are structured, see the Developer Guide.
  6. To make sure that Intel MKL runs on your system, launch an Intel MKL example, as explained in Using Code Examples.

Note: Parts of Intel MKL have Fortran interfaces and data structures, while other parts have C interfaces and C data structures. The Developer Guide contains advice on how to link to Intel MKL with different compilers and from different programming languages.

Get Started
Intel MKL Link Line Advisor