A Heterogeneous Parallel Computing Method for Row Update of Sparse Matrix Lu Decomposition

A sparse matrix and parallel computing technology, applied in the field of sparse matrix LU decomposition, can solve problems such as low efficiency and low operating efficiency, and achieve the effect of improving overall efficiency, avoiding conflicts and dependencies, and increasing computational overhead.

Active Publication Date: 2022-03-29
SHANDONG COMP SCI CENTNAT SUPERCOMP CENT IN JINAN
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In the row update stage, superlu will perform a large number of unit lower triangular matrix equations solving operations, and the efficiency of this type of solving process is not high when calling the blas library provided by the usual TaihuLight, which is much lower than the simple main core The operating efficiency of the , resulting in a computational bottleneck

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Heterogeneous Parallel Computing Method for Row Update of Sparse Matrix Lu Decomposition
  • A Heterogeneous Parallel Computing Method for Row Update of Sparse Matrix Lu Decomposition
  • A Heterogeneous Parallel Computing Method for Row Update of Sparse Matrix Lu Decomposition

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0042] A heterogeneous parallel computing method for sparse matrix LU decomposition row update, including a master core part and a slave core part; such as figure 2 As shown, the specific implementation of the row update main core part is as follows:

[0043] A1) In the row update phase, obtain the row matrix block currently to be processed, analyze each column vector of the row matrix block, count the starting position of each vector in the memory and the size of the vector, and generate an index array; the vector is in The starting position in the memory corresponds to the storage location of the non-zero metadata of the column vector, and the size of the vector corresponds to the number of non-zero metadata of the column vector; the index number of the index array corresponds to the column vector of the row matrix block; A column vector, index number 2 corresponds to the second column vector, and so on;

[0044] The purpose of analyzing each column vector of the row matri...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a heterogeneous parallel computing method for row update of sparse matrix LU decomposition. Based on the superlu algorithm and combined with the master-slave heterogeneous characteristics of TaihuLight, the present invention enables the superlu solver to transfer large-scale computing tasks to the slave core during the matrix decomposition stage and update the row, and utilize the slave core group Efficient computing and data communication capabilities, improve the computing power of solving ultra-large-scale sparse matrix, and further improve the overall performance of the solution. The present invention carries out the mode of task division according to the data of matrix, divides matrix data block by row, at first opens up a matrix space for each slave core, each slave core is responsible for the solution of several matrix rows, in seeking lower triangular identity matrix In the process of linear equations of this type, each line of data is independent of each other, avoiding conflicts and dependencies in the solution space, and successfully solving the equations.

Description

technical field [0001] The invention relates to a heterogeneous parallel computing method for row update of sparse matrix LU decomposition, and belongs to the technical field of sparse matrix LU decomposition. Background technique [0002] In recent years, my country has made a series of significant progress in fields closely related to electromagnetism. Among them, the numerical calculation of electromagnetic field is playing an increasingly important role because of its significant advantages such as high efficiency, flexibility, and convenience. In the numerical analysis method of electromagnetic field, moment Quantitative method has the characteristics of high theoretical precision. Moment method converts the electromagnetic field operator equation to be solved into a matrix equation. Due to its high theoretical precision, when dealing with electromagnetic problems of complex electric and large systems such as airborne phased array The huge complex number sparse matrix pre...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06F17/16
CPCG06F17/16
Inventor 张赞军田敏曾云辉潘景山杨美红
Owner SHANDONG COMP SCI CENTNAT SUPERCOMP CENT IN JINAN
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap