What are the features of CPU/GPU/NPU/FPGA?

CPU/GPU/NPU/FPGA features

Summary of the characteristics of each chip architecture

[CPU]

70% of the transistors are used to build the Cache, and there are some control units with few calculation units for logic control operations.

[GPU]

Most of the transistors are built into computational units with low computational complexity and are suitable for large-scale parallel computing. Mainly used in big data, background server, image processing.

[NPU]

The neurons are simulated at the circuit level, and the integration of storage and calculation is realized by synaptic weights. One instruction completes the processing of a group of neurons to improve the operation efficiency. Mainly used in communication field, big data, image processing.

[FPGA]

Programmable logic, high computational efficiency, closer to the underlying IO, logically editable through redundant transistors and wiring. In essence, there is no instruction, no shared memory, and the calculation efficiency is higher than CPU and GPU. Mainly used in smartphones, portable mobile devices, and automobiles.

As the most common part of the CPU, the CPU performs different tasks in conjunction with other processors. The GPU is suitable for large-scale data training and matrix convolution operations of the background server in deep learning. NPU and FPGA have great advantages in performance, area and power consumption, which can better accelerate neural network calculation. The FPGA is characterized by the development and use of hardware description language, and the development threshold is higher than GPU and NPU.

It can be said that each processor has its advantages and disadvantages. In different application scenarios, it is necessary to weigh the pros and cons according to the needs and select the appropriate chip.