Loop optimization method with multiple condition discrimination

A technology with multiple conditions and multiple loops, which is applied in the computer field, can solve the problems of inconvenient program parallel debugging and optimization, negative impact growth, and increased model calculation time, etc., to achieve the goal of reducing dependency on multiple conditions, improving performance, and reducing the amount of loops Effect

Inactive Publication Date: 2021-03-26
JIANGNAN INST OF COMPUTING TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

With the increase in the scale of numerical simulation problems, the amount of cycles increases massively, and the negative impact of multiple condition discrimination on computational efficiency increases exponentially, and sometimes even becomes one of the main bottlenecks in numerical simulation of application systems
For example, in the initial calculation of the DSMC algorithm, there are multiple conditional discrimination calculations. When the calculation scale is large, the initialization time of the program operation needs more than two hours, which not only increases the mode calculation time, but also brings great difficulties to the parallel debugging and optimization of the program. big inconvenience

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
  • Loop optimization method with multiple condition discrimination
  • Loop optimization method with multiple condition discrimination
  • Loop optimization method with multiple condition discrimination

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0027] Embodiment: a kind of loop optimization method with multiple condition discrimination, multiple loop refers to more than or equal to 4 heavy loops, multiple condition C X 、C Y 、C Z , ... are logical variables;

[0028] The original method is: in the innermost loop of multiple loops, when C X 、C Y 、C Z 、 When the three conditions are judged to be "true" at the same time, the core calculation is performed, and C X The calculation of the value is only relevant for the first heavy cycle, C Y The calculation of the value is only relevant for the second cycle, C Z Calculation of values ​​is only relevant for the third cycle;

[0029] The inventive method comprises the following steps:

[0030] S1. Use performance analysis tools (such as: vtune, gprof) or printout information to analyze the time hotspot function when the program is running, and find out one or more time-consuming program segments in the time hotspot function;

[0031] S2. If the cycle segment found in...

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 loop optimization method with multiple condition discrimination. The method comprises the following steps: S1, analyzing a time hotspot function during program operation, andfinding out one or more time-consuming program segments in the time hotspot function; S2, if the circulation section found in the step S1 is the circulation section judged by multiple conditions, executing the next step; S3, starting a first cycle; S4, starting a second cycle; S5, repeating the process similar to step S3 or S4; S6, judging the logic value of the innermost layer circulation condition: if the logic value is true, carrying out core calculation of the original method, executing step S6, and if the logic value is false, directly starting the next value calculation of the innermostlayer circulation; S7, repeating the step S6 to complete innermost-layer cyclic calculation; and S8, according to the processing method in step S7, completing corresponding cycles in steps S5, S4 andS3 in sequence, and calculating the whole multi-cycle segment. The performance of large-scale parallel computing is remarkably improved, and the solving efficiency of related problems is improved, sothat the computing operation efficiency of the whole application system is improved, and the method has important significance for numerical simulation of actual engineering projects.

Description

technical field [0001] The invention relates to a cycle optimization method with multi-condition discrimination, which belongs to the technical field of computers. Background technique [0002] In rarefied gas dynamics, the DSMC (Direct Simulation Monte Carlo) method is currently the most widely used numerical simulation method. The high-performance computing software system can accurately simulate the aerodynamics / thermals of the high-altitude rarefied flow of the hypersonic vehicle, and the aircraft Reentry thermal protection design, ballistic control, and plume pollution protection design play an important role. The research and numerical simulation of hypersonic vehicles is a major research direction in the field of aerospace in my country. The DSMC numerical simulation method is currently the most widely used method for numerical simulation of rarefied gas dynamics. Design etc. play an important role. [0003] With the research and development of numerical computing, l...

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): G06F30/15G06F30/28G06F30/25G06F30/23G06F111/08G06F111/10G06F113/08G06F119/14
Inventor 徐金秀郭恒黄则强徐占李芳房田文李中华
Owner JIANGNAN INST OF COMPUTING TECH
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