Full-process Supported Simulation Platform Framework Based on Cloud Computing and HPC Integration
Hao Wang,Jinghua Feng,Lin Wang
DOI: https://doi.org/10.62617/mcb658
2024-01-01
Abstract:Simulation technology is widely used in many fields, it usually involves three processes, (i) pre-process, (ii) analysis solver, and (iii) post-process. Simulation calculations require a large amount of computing resources, and users usually need to use cloud and High-Performance Computing (HPC) systems to complete works. Simulation works are increasingly depending on the capacity of HPC or cloud, for cost reasons, people are more willing to use the services than self-built an HPC or Cloud computing cluster. However, that leads to the isolation of calculations and pre- and post-processing work, adding additional time for data transfer. Moreover, simulation engineers also want to use cloud servers in pre-processing and post-processing, since compared with local workstations, cloud servers have significant advantages in saving hardware investment, remote office collaboration, and data integration management. Therefore, we provide a platform framework based on cloud computing and HPC integration that supports the full process of simulation. Then, we implemented it in Tianhe-1A and Tianhe exascale supercomputers and THCloud environments. Through a city area-level explosion simulation experiment, it was verified that the framework can fully support the whole process of simulation, and effectively reduce the time of simulation work, improving the simulation engineer’s work efficiency. The study shows that the platform provides a feasible solution for full-process simulation. Compared with other platforms, it has the characteristics of full-process, high performance and high security.