Data processing system based on brain-like technology

Abstract

The invention provides a data processing system based on a brain-like technology, which comprises a shared data area and a plurality of brain-like computing units communicated with the shared data area, and is characterized in that each brain-like computing unit consists of a computing core, an execution area, a data area and an on-chip network; and when the calculation core receives target data, the calculation core calls the learning model to analyze and pre-judge the input data, configures corresponding parameters of the neuron model and / or synaptic model required by the calculation core, and processes and stores the target data by using the configured neuron model and / or synaptic model. Target data of the next processing link is obtained, and the calculation core further predicts a model corresponding to the calculation resources needed by the next processing link. The system can realize optimal configuration of dynamic resources.

Description

technical field [0001] The invention relates to a data processing system based on brain-inspired technology, belonging to the technical field of intelligent data processing. Background technique [0002] At present, the general database system DBS (Data Base System, referred to as DBS) is composed of software, hardware, database and data administrators, such as figure 1 shown. Among them, the software mainly includes the operating system, application program and database management system; the hardware is generally composed of high-performance computer servers and network equipment, and the computer server includes chassis, main board, memory, hard disk, network card, power supply and other components. function and data support carrier; the database is managed uniformly by the database management system, and data insertion, modification and retrieval must be performed through the database management system; the data administrator is responsible for creating, monitoring and ...

AI Technical Summary

Problems solved by technology

However, the separation of storage and calculation still requires the system to continuously read data from the memory and send it to the processor. There is a lot of energy loss in this process, and the faster the speed, the higher the power consumption.

[0005] (2) Database management programs and algorithms rely on manual labor: currently, the processing of data in general database systems completely depends on administrators, or the fixed algorithms and processing mechanisms of management programs. In special occasions, administrators cannot handle them in time, or algorithms and processing mechanisms Events that are not considered may lead to failure or abnormal conditions, which greatly limits the usage environment or applicable scenarios of general database systems

[0006] (3) High memory usage rate and low execution efficiency: The servers used in general database systems are generally equipped with high-speed and large memory. After the system is started, the database or database index will be transferred into the memory; The search program will be executed, and the data in the memory will be read frequently; when the search task is too heavy, the processor and memory resources will be almost completely occupied, and the system will be seriously stuck or even shut down

Images