Study On Lattice Boltzmann Method Of Biofilm Materials And GPU Parallel Computing

The basic idea of parallel computing is to decompose the computing problem into several parts,and then call multiple computing resources to calculate collaboratively.The traditional central processing unit(CPU)parallel computing system consists of two or more computers,and the computers transmit data through network switch.The disadvantage is that the total amount of data will significantly reduce the transmission speed.In a graphics processing unit(GPU)based on the multiple-core architecture,data can be transferred between cores through the PCI-E interconnect switch with higher transfer efficiency.Excellent computing power is also inseparable from an excellent programming platform.NVIDIA has added an easy-to-use programming interface to the GPU-Unified Computing Architecture(CUDA).It is extended on the basis of standard C,so that programmers can specify parallel programming model writing,which greatly reduces the threshold for programming using GPU.Lattice Boltzmann Method(LBM)is a mesoscopic model between the microscopic molecular dynamics model and the macroscopic continuous model.It uses a distribution function to replace the traditional numerical method to solve the macroscopic variables.Simplify the difficulty of fluid calculation.LBM has the advantages of clear physical background,low programming difficulty,and simple boundary processing,so it is very suitable for GPU parallel computing.The lymphatic system plays an important part in the body's immunity and cell's internal environment homeostasis.Similar to the blood circulatory system,the lymphatic system is also a piping system throughout the body,which is mainly composed of lymphatic fluid and lymphatic vessels.The spontaneous contraction of the lymphatic vessels drives the flow of lymphatic fluid in the vessels.The contraction-relaxation mechanism of lymphatic vessels is determined by the oscillating feedback loop of Ca²⁺and NO concentration.The production of NO depends on the shear force acting on the lymphatic valve.In this paper,the lymphatic valve is used as the biofilm material.In the real lymphatic system,the distribution of NO in a certain section of lymphatic vessels will be affected by the lymphatic vessels connected to it,especially the upstream connecting fragments.In this paper,LBM model with multi-segment lymphatic vessels as the physical background is established through the CUDA programming platform,and the contraction of multi-segment lymph vessels and their internal physical properties are studied through this model.The research content of this article includes the following aspects:1.Introduce the research history and basic principles of LBM,and verify the method based on theory.The boundary processing conditions,dimension conversion and numerical realization process of D2Q9 model in LBM were studied.2.Using the CUDA programming platform to implement the GPU parallel computing of the LBM Poisson flow model,try to reduce the latency of accessing the memory by optimizing the addressing method,and then achieve the purpose of improving the computing efficiency.Under the premise of ensuring the correct calculation results,the speedup ratio of71 times and 25 times that of the CPU was obtained on the GPU through linear addressing and subscript addressing respectively.On the basis of the subscript addressing program,the function parameters are simplified by using the structure,and the structure is put into the shared memory for accelerated optimization.3.Based on the D2Q9 model,a multi-segment lymphatic vessel model containing three types of lymphatic vessels was established.The calculation parameters of lymphatic vessels can be adjusted by code modification.The systolic-diastolic cycle of lymphatic vessels was reproduced by this model,the distribution of NO and pressure in multi-segment lymphatics and the change of average NO concentration,the flow of each tube in the three-segment tube model changed with time were studied.