Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Optimal dimension solving method and apparatus of kernel function in GPU programming model

A technology of optimal size and programming model, applied in the direction of processor architecture/configuration, details related to image processing hardware, etc., can solve problems such as unreasonable division of GPU high performance and impact

Inactive Publication Date: 2017-05-10
CHINA AGRI UNIV
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The present invention provides a method and device for solving the optimal size of the kernel function in the GPU programming model, which is used to solve the problem that the unreasonable division of the size of the kernel function in the prior art affects the high performance of the GPU

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
  • Optimal dimension solving method and apparatus of kernel function in GPU programming model
  • Optimal dimension solving method and apparatus of kernel function in GPU programming model
  • Optimal dimension solving method and apparatus of kernel function in GPU programming model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0084] The specific implementation manners of the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0085] figure 1 A method for solving the optimal size of the kernel function in a GPU programming model of the present invention is shown, including:

[0086] S11. Obtain the structural parameters of the GPU hardware structure;

[0087] S12. Obtain the analysis parameters preset in the parallel application program;

[0088] S13. Establish a time cost model according to the structural parameters and the analysis parameters;

[0089] S14. Establish a size optimization model according to the time cost model;

[0090] S15. Solve the size optimization model to obtain the optimal size of the kernel function and output it.

[0091] Each step of the above method is explained ...

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 provides an optimal dimension solving method and apparatus of a kernel function in a GPU programming model. Basic requirements of parallel application programs can be met by constructing a time cost model; effectiveness of time cost prediction can be ensured; and the prediction result becomes real. The analysis on the kernel function dimension becomes comprehensive and reasonable and by analyzing parameters in the time cost model, adding constraint conditions, and constructing a dimension optimization model, so that the model precision and operation efficiency can be improved. On the basis of the classified discussion ideal, three kinds of ideas that may occur during parallel program execution on a GPU are analyzed in detail and the kernel function dimension is calculated under each circumstance; and a simple formula that is easy to operate is obtained to guide a user to carry out rapid dimension selection. Under the circumstance that the parallel application program is not operated, a static analysis is carried out to obtain an optimal dimension of a kernel function, so that the work and production efficiency can be improved substantially.

Description

technical field [0001] The invention relates to the field of parallel and distributed computing processing, in particular to a method and device for solving the optimal size of a kernel function in a GPU programming model. Background technique [0002] Since the beginning of this century, graphics processing units (GPUs) have gradually begun to evolve from specific graphics accelerators to general-purpose computing devices. In recent years, GPU has been designed to efficiently process computer images, and its computing power has also been greatly improved, and it has higher computing power than CPU. Therefore, more and more users begin to use high-performance GPUs to solve complex application problems. [0003] In addition, time overhead is the most basic measurement method for parallel algorithms, and other evaluation criteria should be aimed at reducing time overhead. The research on the time cost equation of parallel computing is mainly to combine the characteristics of...

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
IPC IPC(8): G06T1/20
CPCG06T1/20G06T2200/28
Inventor 王剑秦王佳弘林枝宏孙瑞志
Owner CHINA AGRI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com