Step #2 concerned implementing some optimizations in order to increase the overall frame rate.
As usual, before implementing any optimization, a profiling the code was carried out profiled in order to detect the portion of code portions that made sense to optimize. In addition to traditional, well-know techniques, the specific NPU-related tools were used as well. For instance, the following dump shows the a detailed report referring to the execution of a Convolutional Neural Network (CNN) on the accelerator.
{| class="wikitable"