Parallel Computing And Optimization Of Large-scale Particles Based On GPU

In the recent years discrete element model (DEM) simulations have been increasingly used to study and analyze flows of particulate systems in the interaction between particles High performance and efficiency for parallel computing has an important significance in large scale discrete element method (DEM) simulation. After analyzing a simulation framework of DEM built on the Graphic Processor Unit (GPU) platform with CUDA architecture and evaluating the simulated data, we propose three optimization methods to improve the performance of the system.We apply stencil computation model to the particle searching and calculation of forces based on gridding to formulate the structure in the particle-particle contact and neighboring particle searching for parallel computing. Stencil computation model is crucial to the core of many scientific computing applications since it is among the most important and time-consuming kernels in many scientific applications; After analyzing hardware properties and combining with the characteristics of the example, we seek out a reasonable and effective parallel granularity by altering the number of blocks and threads in the code. The parallel granularity is one of the key problem in the parallel processing system for utilizing multi-core platforms. We set up a shared-memory environment for data prefetching and storing the results of intermediate calculations by rational analysis and calculation, because shared-memory could achieve better load balancing for the calculation of particle models.Dilated polyhedral element can accurately construct geometric shape of irregular particles, which can improve the accuracy of contact-impact computing remark-ably. This framework completes parallel computing acceleration of large-scale dilated polyhedral particles based on GPU by using CUDA. Loose coupling among different parts of computing framework enable it to support different mechanical models. From the perspective of user and adequate consideration of the characteristics of DEM calculation, UI interface which is implemented under Qt platform is de-signed elaborately. OpenGL is used for visualization of calculation results. Tests of multiple sets of data show that this framework can effectively simulate and display scenarios of complex particles contact-impact process dynamically.