This KB article explains the different changes done in the OpenMP support in Intel Performance libraries (Intel® IPP and Intel® MKL).
The article describes serveral possible scenarios and solution for high cpu usage when call IPP function in real application
The parallel direct sparse solver in Intel® MKL called PARDISO has many parameters that must be understood and properly set to get the expected results. These tables may help by putting some of the required information in a more readable format.
A technical talk on the Condition Numerical Reproducibility (CNR) feature in Intel® MKL 11.0
A tutorial on how to use #pragma simd and SIMD-enabled function features in Intel® Cilk™ Plus.