FlowModellium Laboratory (MIPT): Past, Current & Future Activity. Near-wall domain decomposition method.
讲座名称：
FlowModellium Laboratory (MIPT): Past, Current & Future Activity. Near-wall domain decomposition method.
讲座时间：
2018-09-05
讲座人：
Mikhail Petrov
形式：
校区：
兴庆校区
实践学分：
讲座内容：
时间： 2018年9月5日上午10:30-11:30
地点： 航天航空学院教一楼第四会议室
报告题目：FlowModellium Laboratory (MIPT): Past, Current & Future Activity. Near-wall domain decomposition method.
报告人：Mikhail Petrov博士
报告摘要：
The presentation will be devoted to FlowModellium Laboratory (MIPT) and the research is carried out there. The author will pay attention to one of the research areas: near-wall domain decomposition method. A near-wall domain decomposition method for use in turbulence modelling is applied to the k-ω (SST) and Spalart-Allmaras turbulence models. The near-wall region is excluded from the main computational mesh. This eliminates the expense of computing the solution in the viscous sub layer and the total computation time.
报告题目：Parallel computational methods for high-speed aerodynamics on multi/many core architectures
报告人：Alexander Chikitkin博士
报告摘要：
Modern supercomputers utilizies processors with large number of cores, up to 72 as for Intel Xeon Phi generation. At the same time, power and memory of one core in such processors is much less compared to processors with moderate number of cores (6-8). That is why it is crucial that numerical algorithm has good strong scalability on such computers.This requirement is a big challenge for implicit schemes due to strong data dependence. Implicit schemes are anavoidable in numerical simulations of wide range of phenomena such as multi-phase flows, reacting high-speed flows because of numerical stiffness.We describe parallel computational methods for high-speed aerodynamics implemented in research code "Flow modellium" and present scalability results obtained on modern supercomputers with up to 60000 threads. Proper modification of parallel implicit scheme algorithm allows to achieve satisfactory parallel efficiency (70-80%).
相关视频