Intel(R) Math Kernel Library for Deep Neural Networks (Intel(R) MKL-DNN)  1.0.4
Performance library for Deep Learning
Benchmarking Performance

Intel(R) Math Kernel Library for Deep Neural Networks (Intel(R) MKL-DNN) has a built-in benchmarking program call benchdnn.

For a complete description of the available options and working examples, see the https://github.com/intel/mkl-dnn/blob/master/tests/benchdnn/README.md "benchdnn readme".