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

Graphics processing unit (GPU) oriented bitonic merge sort method

A sorting method and vector technology, which can be used in multi-programming devices, concurrent instruction execution, resource allocation, etc., and can solve problems such as affecting computing efficiency.

Active Publication Date: 2012-10-24
COMP NETWORK INFORMATION CENT CHINESE ACADEMY OF SCI +1
View PDF0 Cites 24 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Its dual-tone sorting program can make full use of the GPU's stream processor, but the synchronization work in the sorting process is completely completed by the CPU, and thread synchronization between work groups requires context switching, which affects computing efficiency

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
  • Graphics processing unit (GPU) oriented bitonic merge sort method
  • Graphics processing unit (GPU) oriented bitonic merge sort method
  • Graphics processing unit (GPU) oriented bitonic merge sort method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0024] The technical solutions of the present invention will be described in further detail below with reference to the accompanying drawings and embodiments.

[0025] The present invention mainly improves the GPU double-tone merge and sort method in the prior art through the following methods:

[0026] 1. Simulate multiple comparators using vectors

[0027] In the traditional GPU doubly merge sort method, one thread acts as a comparator (compare and conditionally interchange), and the length of the column to be sorted is twice the number of comparators. The comparator can be grouped for the first time, and the group number can be used to determine whether the arranged data segments are arranged in ascending or descending order. At the same time, the comparator can be grouped twice, and the position of the sequence element operated by the comparator can be obtained through the group number , figure 1 It is a simple schematic diagram of a double tone merge sorting network wit...

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 graphics processing unit (GPU) oriented bitonic sort method and a system. A plurality of comparators are simulated through one thread operation vector, a plurality of threads are subjected to a merge sort in a parallel execution mode, the merge sort of threads in the same workgroup is completed by using synchronous functions in a synchronous mode, the merge sort of threads in different workgroups is completed through a central processing unit (CPU) in the synchronous mode, further, a plurality of vectors can be used to simulate vectors which are much longer, during the sort process, write back addresses of operands of the comparators are changed, and the memory read / writes are optimized. According to the graphics processing unit (GPU) oriented bitonic merge sort method, on a basis of saving the storage space, synchronous times between the CPU and the GPU are effectively reduced, the total quantity and delayed time of executing instructions are reduced, and the use ratio of a GPU computing unit is increased.

Description

[0001] The invention relates to a data sorting method, in particular to a GPU-oriented dual tone merge sorting method based on OpenCL specification. Background technique [0002] Sorting is one of the most common operations in computer applications. With the further development of parallel processing technology, parallel sorting has become a very important research field. Parallel sorting is usually divided into two categories: one is direct sorting, which can directly realize the sorting of sequences; the other is merge sorting, which can quickly merge multiple sorted sequences into one sorted sequence. [0003] In the prior art, most sorting methods need to open up a new memory space to store the results of the intermediate steps of sorting, such as common quick sorting, radix sorting, and bucket sorting in parallel sorting algorithms. The double tone merge sorting method can directly exchange data in the storage space of the column to be sorted, which effectively saves memo...

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): G06F9/38G06F9/50
Inventor 迟学斌王珏阚圣哲聂宁明郎显宇
Owner COMP NETWORK INFORMATION CENT CHINESE ACADEMY OF SCI
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