Tips to measure the performance of Intel® MKL with small matrix sizes

The time required by the first Intel® MKL call should be ignored for the performance measurements. The first Intel MKL call has overhead due to buffer allocation and thread initialization. Ignoring the first Intel MKL call gives more consistent times for small problems.
Authored by Ying H. (Intel) Last updated on 10/26/2017 - 22:33
For more complete information about compiler optimizations, see our Optimization Notice.