Scheduling method of reconfigurable computing architecture based on LU decomposition of arbitrary dimension matrix

Abstract

The invention belongs to the field of matrix computing technology, and particularly relates to a scheduling method of a reconfigurable computing architecture based on LU decomposition of an arbitrary dimension matrix. The reconfigurable computing architecture based on LU decomposition of the arbitrary dimension matrix consists of processing units, an interface controller, a scheduling module and a storage module. The method comprises the following steps: interconnecting the processing units, and forming a reconfigurable processing array with the interface controller; enabling the scheduling module to design a scheduling mechanism, generating configuration information, and issuing the configuration information to the reconfigurable processing array; constructing a computing architecture that meets the LU decomposition of a current dimension matrix; based on the constructed computing architecture, distributing computing data to the processing units according to the scheduling mechanism, and performing LU decomposition computation; and returning an LU decomposition computation result to the storage module through the interface controller. According to the scheduling method disclosed by the invention, the LU decomposition computation of the arbitrary dimension matrix can be achieved through the mode that the fixed processing units can be reconfigured, and the flexibility of LU decomposition can be improved.

Description

technical field [0001] The invention belongs to the technical field of matrix calculation, and in particular relates to a scheduling method of a reconfigurable calculation structure based on LU decomposition of an arbitrary-dimensional matrix. Background technique [0002] LU decomposition is a matrix decomposition method in scientific computing. It is often used in the solution of linear equations, matrix inversion and determinant calculation. It is also widely used in many fields such as graphics and image processing and scientific computing. The LU decomposition method is to decompose the original square matrix into an upper triangular matrix and a lower triangular matrix. [0003] When the matrix is ​​a non-singular matrix of order N and all sequential principal subforms are not 0, the matrix can be decomposed into a product of a lower triangular matrix L with all 1s in the main diagonal elements and an upper triangular matrix U, that is, A=LU, and Decomposition is uniq...

AI Technical Summary

Problems solved by technology

[0004] At present, most matrix operations are implemented based on software, but as the dimension of the matrix increases, the speed of software processing will be greatly slowed down.

The LU decomposition of large-dimensional matrix based on hardware can give full play to the speed of hardware, but the LU decomposition of large-dimensional matrix requires high computing resources and storage space. The existing parallel structure design of high-performance matrix LU decomposition needs to compare the algorithm. Substantial parallelization rewrites or constructs complex parallel models, and implements the corresponding proprietary parallel structure, which consumes more resources, but the reusability and scalability are not ideal, and it is difficult to realize the LU decomposition of variable-dimensional matrices However, the design and execution time complexity of using a more general linear array to realize LU decomposition is high, and the data parallelism in LU decomposition calculation cannot be fully exploited; the LU decomposition method based on block matrix calculation can realize the matrix of any dimension Decomposition calculation, but the current implementation method does not make full use of the inherent line and streamlined features of decomposition calculation, which makes it difficult to improve the calculation efficiency

Images