Large-scale matrix QR decomposition parallel computing structure

Abstract

The invention discloses a large-scale matrix QR decomposition parallel computing structure, and relates to the field of digital signal processing. The invention aims to provide a three-level parallelcomputing structure with high throughput rate and low delay. The invention is realized through the following technical scheme: a multi-core processor chip is adopted to construct a processor cluster system and a QR decomposition parallel computing structure. The top architecture divides a matrix to be decomposed into a plurality of data fragments, the data fragments are distributed to all levels of nodes through an interconnected communication network among multi-core processor nodes, all the levels of nodes are sequentially calculated step by step according to a binary tree complete structure, and all the levels of nodes are subjected to parallel calculation; the middle-layer architecture performs matrix partitioning, and performs operation layer by layer along diagonal sub-arrays; the underlying architecture uses a processor instruction set to perform multi-data parallel vector calculation to complete single-core QR decomposition and multiplication operations. And the multi-core processor cluster adopts a layer-by-layer decomposition structure to realize QR parallel decomposition of a large-scale matrix.

Description

technical field [0001] The invention relates to a large-scale array antenna signal processing technology in the field of digital signal processing, in particular to a large-scale multi-core processor cluster parallel structure QR decomposition method for high-performance parallel computing in the field of numerical calculation by a large-scale parallel processor. Background technique [0002] In the field of digital signal processing, signal processing algorithms such as large-scale array antenna signal processing and large-scale multiple-input multiple-output (MIMO) technology often involve issues such as covariance matrix inversion, channel matrix estimation, and channel equalization. QR decomposition has great significance in these aspects. widely used. Among them, MIMO is a standard for wireless network evolution. In fact, MIMO technology has become one of the most critical technologies in many wireless communication standards, such as IEEE802.11n, 3GPP-LTE and other st...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to provide a kind of clear parallel logic, strong scalability and portability, high throughput, low delay and high versatility for large-scale multi-core processor clusters involving large-scale QR decomposition and requiring real-time processing. It can make full use of the parallel processing advantages of multi-core processor clusters to realize the three-level parallel computing structure of multi-processor node parallelism, single-processor multi-core parallelism and single-core multi-data parallelism, so as to solve the problem that the traditional QR decomposition method cannot effectively utilize multi-core processor clusters resources for massively parallel computing

Images