Intel MKL

Using Intel MKL on SeaWulf

Overview

The Intel Math Kernel Library (MKL) is a highly optimized library for performing mathematical operations such as matrix multiplication, Fast Fourier Transforms (FFT), and vector operations. On SeaWulf, Intel MKL is available in various versions for different needs, enabling users to choose the best fit for their application.

Available Modules on SeaWulf

Below are the available Intel MKL modules for Haswell nodes:

Haswell Nodes

  • intel/mkl/64/2017/0.098
  • intel/mkl/64/2018/18.0.0
  • intel/mkl/64/2018/18.0.1
  • intel/mkl/64/2018/18.0.2
  • intel/mkl/64/2018/18.0.3
  • intel/mkl/64/2019/19.0.0
  • intel/mkl/64/2019/19.0.3
  • intel/mkl/64/2019/19.0.4
  • intel/mkl/64/2020/20.0.0
  • intel/mkl/64/2020/20.0.1
  • intel/mkl/64/2020/20.0.2

Loading a Module

To use the Intel MKL module, load the appropriate version based on your application requirements. You can load the module using the module load command. For example:

module load intel/mkl/64/2020/20.0.2

Make sure to load the version that fits your needs for optimal performance.

Using Intel MKL

Once the Intel MKL module is loaded, you can link your program with MKL to accelerate mathematical computations. For example, here’s how you might compile a program that uses Intel MKL:

icc -o my_program my_program.c -lmkl

Ensure that you link your program with the -lmkl flag to access Intel MKL’s functionalities.

Important Notes

Intel MKL is optimized for high-performance computing and supports multi-threading, which can improve performance on parallel systems. Choose the version of MKL that is most compatible with your application and environment.

Incompatible versions of Intel MKL or incorrect compiler flags may lead to performance degradation or errors. Always ensure that you use the appropriate version and compilation flags for your application.

Additional Resources

If you need further assistance with Intel MKL or have any questions, please refer to the SeaWulf documentation or contact support for help with optimizing your use of Intel MKL.