What is CPU,GPU and TPU? Understanding these 3 processing units using Artificial Neural Networks.

In this video, we will be explaining the difference between Central Processing Unit (CPU), Graphics Processing Unit (GPU) and Tensor Processing Unit (TPU).

Neural Network is all about Matrix Operations and how each of the processing units handles these matrix multiplications and additions.
CPU is a General Purpose Process Architecture. It handles the matrix operation in a scalar way, sequentially one by one. It can handle tens of operations per cycle.

GPU is also a General Purpose Process Architecture, but it handles the matrix operation in the form of vector way. It can handle tens of thousands of operation per cycle.

TPU is a domain specific built architecture. Google designed it as a matrix processor specialized for neural network loads. It handles the matrix operation in the form of Tensor and it can handle up to hundreds of thousands of operations per cycle.

#CPU #GPU #TPU #Artificial Neural Networks #MNIST #Machine Learning #Deep Learning