Font Size: a A A

Research Of Multi-core CPU Parallel Computing Of 2D Incompressible Laminar Flow Problems Based On EFG Method

Posted on:2019-06-19Degree:MasterType:Thesis
Country:ChinaCandidate:H WuFull Text:PDF
GTID:2428330548982034Subject:Master of Engineering
Abstract/Summary:PDF Full Text Request
The Element-Free Galerkin Method(EFG)has the advantages of simple pre-processing,high accuracy,and rapid convergence.It has been widely used in the field of structural solid mechanics,but its computational complexity is large,computational time-consuming,etc.Disadvantages also hinder its promotion to other fields.At the same time,because multi-core CPUs have excellent computing power,they have been widely used in scientific research,engineering calculations,and disaster prevention and mitigation.In this paper,the EFG method is applied to the fluid flow problem,and its multi-core CPU parallel computing method is studied.The main contents of this paper are as follows:(1)This paper starts with the relevant theories and formulae of the EFG method applied to the two-dimensional laminar flow problem,and uses the region search method and the cross-node pair method to search the computational domain nodes and assemble the stiffness matrix,and constructs a layer applicable to the layer.An improved EFG algorithm for the parallel computation of flow problems was verified by plate extrusion and 900 elbow examples to verify the feasibility of the proposed algorithm.(2)Researched and analyzed the design principle of multi-core CPU parallel algorithm.Based on task parallelism and data parallel technology,parallel optimization of the proposed improved algorithm was designed.The parallel,velocity-pressure matrix of the search function algorithm in the flow field was included.The multi-core CPU parallel computing flow was obtained through the storage and assembly and the parallelization of the equations.The multi-core CPU parallel computing was verified by a numerical example,and the calculated results were compared with the serial calculation..(3)Introduce performance indicators for measuring parallelization characteristics,perform single-core performance analysis,and select slabs with different orders of magnitude,call different numbers of CPU cores,set different number of workers,etc.for flat plate extrusion and 900 elbows,etc.The examples are based on parallel computing of multi-core CPUs.The relationship between the influencing factors such as the number of nodes,the number of called cores,and the number of workers and the speedup ratio is studied and analyzed.In summary,this paper applies the EFG method to the numerical method of the two-dimensional laminar flow problem and establishes its parallel computing flow using multi-core CPU to improve the computational efficiency of the EFG method in solving fluid flow problems.The research results obtained in this paper are based on multi-core.The application of EFG method under CPU hardware provides a good theoretical reference value.
Keywords/Search Tags:Two-dimensional laminar flow, EFG method, multi-core CPU, speedup, parallel computing
PDF Full Text Request
Related items