I would like to discuss about some performance data of latest version of Intel Math Kernel Library (Intel MKL 10.2) functions.
There is a PDF file in the MKL web page (http://software.intel.com/sites/products/collateral/hpc/mkl/charts.pdf), which gives you performance data for the following domains/functions.
BLAS (DGEMM) - The BLAS and LAPACK functionality included in Intel MKL is highly optimized for Intel processors and can significantly increase the performance of your application compared to alternative
implementations.
Fast Fourier Transforms (FFT) - The Fast Fourier Transform functionality of the Intel MKL offers outstanding performance and scalability across a wide range of problem sizes for both SMP and distributed memory (cluster) systems.
ScaLAPACK - The Intel MKL implementation of ScaLAPACK is highly optimized for Intel processors and can significantly increase the performance of your application compared to other implementations of ScaLAPACK.
Vector Math - The VML included with Intel MKL provides highly optimized vector implementations of computationally intensive core mathematical functions.
All these performance numbers are measured on Quad-Core Intel Xeon Processor W5580 3,2Ghz, 8 MB L2 cache, 12 GB Memory.
