Deep learning operation - is there any formal definition?


In the announcement of the new self-driving car platform Drive PX2 there is a piece of information about the performance of the system, namely, 24 Trillion Deep Learning Operations per Second. Is there any formal definition of “deep learning operation”? I’m currently running some calculations related to deep learning and it would be useful to have some sort of unit to calculate in (if it is a real “thing”).


P.S.: Sorry if I’m at the wrong topic, I couldn’t find a closer one.
The announcement is here:

You may refer below NVIDIA cuDNN – GPU Accelerated Deep Learning page as a start.


Watch the demo of deep learning using eSOMTK1 - NVIDIA Tegra K1 System on Module. Deeplearning is the set of algorithms to train the model data and based on trained data properties, we can recognize objects. Has lot of potential in autonomous driving systems - driver-less cars.

eSOMTK1 is based on NVIDIA Tegra® K1 CPU in a compact form factor. This Tegra K1 System on Module has NVIDIA Tegra® K1 4-Plus-1™ ARM® Cortex™-A15 Quad core running at 2.2GHz with DDR3L SDRAM configurable up to 4GB, eMMC configurable up to 64GB.

Buy Now :