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

Rhombus partition parallel optimization method for template calculation

An optimization method and rhombus technology, applied in computing, concurrent instruction execution, machine execution devices, etc.

Active Publication Date: 2020-07-07
XI AN JIAOTONG UNIV
View PDF6 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] With the iterative update of computing devices, the "storage wall" problem caused by the imbalance between the computing power of computing devices and the communication bandwidth of the system is still a bottleneck problem in improving program performance.

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
  • Rhombus partition parallel optimization method for template calculation
  • Rhombus partition parallel optimization method for template calculation
  • Rhombus partition parallel optimization method for template calculation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0043] The present invention is a kind of parallel optimization method of rhombus block for template calculation, comprising the following steps:

[0044] S1. Diamond block strategy for template calculation

[0045] S101. Define template calculation

[0046] Template computing is actually a method of completely traversing all spatial grid points under a large number of time step iterations in a fixed pattern on a structured spatial computing grid, and a certain fixed pattern used to update the grid points is called a template.

[0047] A template is to define how the value of a point in space is calculated or updated from itself, the values ​​of adjacent points, and the values ​​of adjacent points at different time steps.

[0048] Jacobi is a typical template calculation. The core code of Jacobi-1d is as follows:

[0049]

[0050]

[0051] The fixed template for data update calculated by the above template is represented by the following formula:

[0052]

[0053] ...

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 rhombus partition parallel optimization method for template calculation. The rhombus partition parallel optimization method comprises the following steps of: defining a d-dimensional space iterative domain calculation definition template, selecting a rhombus partition shape and a Cache size of a computer system in which template calculation runs according to a calculatedslope of a template and the width and height of rhombus partitions, determining the size of the rhombus partitions, and determining a rhombus partition strategy for template calculation; and performing a parallel scheduling strategy on the obtained rhombic partitions, adopting a wavefront parallel strategy on a certain dimension, which is not subjected to rhombic partition division, of the rhombicpartitions, and adopting a coarse-grained parallel strategy of all rhombic partitions and a fine-grained wavefront parallel strategy in the rhombic partitions for template calculation to complete parallel optimization. According to the rhombus partition parallel optimization method, compared with an original template calculation program, the optimized template calculation program can obtain better data locality and better program parallelism, so that the running time is further shortened.

Description

technical field [0001] The invention belongs to the technical field of locality and parallel development of computer programs, and in particular relates to a parallel optimization method for diamond-shaped blocks for template calculation. Background technique [0002] Stencil Computations, based on the iteration of fixed templates and new data, have attracted attention as the core of computing in the fields of image and video processing, large-scale scientific and engineering simulations, and computing. Common template calculations include structured grid calculations in computational fluid dynamics (LBM, etc.), explicit and implicit partial differential equation solving (PDEsolver), Jacobi (Jacobi) and Gauss-Seidel (Gauss-Seidel) methods And the finite difference time domain method (FDTD) and so on. [0003] With the iterative update of computing devices, the "storage wall" problem formed by the imbalance between computing power of computing devices and system communicatio...

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): G06F9/38
CPCG06F9/3836G06F9/3856
Inventor 伍卫国王倩刘松屈彬
Owner XI AN JIAOTONG 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