Method for positioning performance problems of large-scale parallel program

An implementation method and large-scale technology, applied in the field of parallel program performance analysis, can solve problems such as difficulty in automatically locating program performance, and achieve the effect of reducing complexity

Inactive Publication Date: 2011-05-18
BEIHANG UNIV
View PDF2 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] At present, most parallel program performance analysis tools use dynamic analysis methods. Through the visual display of program running data, the performance characteristics of program running are manually checked, and it is difficult to automatically locate the performance bottleneck of the program.

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
  • Method for positioning performance problems of large-scale parallel program
  • Method for positioning performance problems of large-scale parallel program
  • Method for positioning performance problems of large-scale parallel program

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0039] This method is a large-scale parallel program performance problem location implementation method. Firstly, the functions in the parallel program are grouped, and then the causality diagram of the parallel program performance problem is established, and then the function performance data set recorded in the program log is clustered. Calculation, according to the calculation results, select the expensive function group, input it into the classifier, and realize the positioning of the performance problem through the mapping of the function group to the performance problem. The parallel program performance problem location realization system of the method mainly includes a module for grouping functions, a module for establishing a causal relationship diagram, a function performance data clustering calculation module, and a performance problem location module. Hereinafter, a typical matrix multiplication MPI parallel program is taken as an example to describe the method in de...

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 discloses a method for positioning performance problems of a large-scale parallel program, belonging to the technical field of parallel program performance analysis. The method comprises the following steps of: grouping functions in a parallel program; establishing a cause-effect relationship diagram; performing clustering computation on function performance data of profile logs and track logs in the parallel program; selecting function groups with large overhead according to the clustering computation results and inputting the function groups into a sorter; and performing performance problem positioning according to the mapping from the function groups to performance problems. With the method disclosed in the invention, the performance problems of the parallel program can be accurately positioned and results can be interpretable.

Description

technical field [0001] The invention relates to the technical field of parallel program performance analysis. Background technique [0002] At present, supercomputers are large-scale parallel systems, and with the development of computer architecture and hardware, the peak performance of supercomputers is getting higher and higher, but the performance of parallel programs has not been improved accordingly. Due to the large scale and high complexity of parallel programs, it is difficult for programmers to find performance problems in program execution, resulting in low utilization of parallel systems. [0003] Therefore, the performance analysis of parallel programs through automated performance analysis tools can provide reasonable suggestions for users to find performance problems, locate performance bottlenecks and propose performance improvement measures. At the same time, by providing a multi-dimensional view of performance data, users can manually analyze the data of i...

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 Applications(China)
IPC IPC(8): G06F11/36
Inventor 李巍李敏李云春朱鹏
Owner BEIHANG UNIV
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