Electric power intelligent control terminal and SOC electric power chip architecture based on RISC-V

Abstract

The invention relates to an electric power intelligent control terminal and an SOC electric power chip architecture based on RISC-V. The SOC electric power chip architecture comprises an RISC-V main control unit, an AI engine unit, a peripheral unit, a memory and a power supply unit which are integrated on a chip and interconnected through a bus. The RISC-V main control unit is used for processinga control instruction and a control instruction execution process; various AI operators are arranged in the AI engine unit and used for executing an AI algorithm according to the control instruction;the peripheral unit is used for providing a universal interface for accessing peripheral equipment into the chip; the memory is used for storing control instructions; and the power supply unit is used for supplying power to the chip. Through efficient interaction between the built-in RISC-V main control unit and the AI engine unit, on one hand, open source autonomy is achieved based on an open source instruction set of the RISC-V, the safety is improved, on the other hand, expandability of the RISC-V instruction is used for achieving expandability of the AI instruction, in cooperation with various built-in AI operators in the AI engine unit, rapid execution of an AI algorithm in the AI engine unit is achieved, and the operational performance of a chip is improved.

Description

technical field [0001] The invention relates to an intelligent control terminal of a power system, in particular to a power intelligent control terminal and its RISC-V-based SOC power chip architecture. Background technique [0002] With the deepening application of artificial intelligence technology in the power field, cloud-device collaboration will become a widely deployed computing paradigm. Compared with cloud training, terminals often perform intelligent inference calculations, and intelligent algorithms sink to terminal devices to provide intelligent inference edge computing (Edge Computing), which can meet the requirements of real-time performance, low power consumption, and security, and save bandwidth and storage. The RISC-V industry has just started in China, and it is very suitable for application in terminal intelligent reasoning and the development of artificial intelligence chips. [0003] At present, the multi-core heterogeneous architecture has become a ty...

AI Technical Summary

Problems solved by technology

However, there is no mature structure for the power industry, which cannot meet the requirements of independent property rights, safety and controllability of the power grid, and does not match the needs of the power grid. Independent Customization Requirements for IoT Terminal Computing

Images