HPC software and applications
CNIC has more than 40 self-developed high performance computing (HPC) software and over 140 HPC applications.
Examples of HPC applications
CNIC's HPC software, tools and applications have played an important role in supporting research innovations and serving societal needs with several recent examples as follows.
CNIC's HPC team developed the large-scale paralleled computational fluid dynamics software system for aircraft design named China Computational Fluid Dynamics (CCFD) which includes multi-block structured solver module CCFD-MGMB, multi-body separation simulation module CCFD-MBS and aeroelasity simulation module CCFD-AE. Capable of simulating true aircraft shapes with more than 10000 processors, CCFD has been used by more than ten domestic aircraft design organizations or aerodynamics research institutions such as the First Aircraft Institute of Aviation Industry Corporation of China (AVIC), Shenyang Aircraft Institute of AVIC ,China Sanjiang Astronautics Group, Beihang University and Northwestern Polytechnical University by the end of 2015, making remarkable contributions to the development of aeronautics and astronautics in China.
In 2015, CNIC's HPC team assisted Beijing Public Security Bureau to provide direct clues for more than 240 cases using the meta supercomputer and effectively supported the public security of the capital city. Various systems have been developed based on the close cooperation with researchers in Beijing Public Security Bureau and in-depth understanding of criminal investigation workflows. For instance, the automated fingerprint identification system based on volunteer computing is developed to solve the shortage of computing resource in Beijing Public Security Bureau using the idle computing resources such as personal computers. The large-scale real-time fingerprint query system is developed to achieve on-site task submission and relevant results feedback using the interfaces with the national grid and effectively circumventing sensitive personnel information and features. The fingerprint query system can feed back relevant results within an hour and facilitate near real-time criminal investigation and detection. In addition, palmprint recognition algorithm based on GPU broke the speed bottleneck of palmprint recognition, significantly increasing the detection means of the investigators.
By the end of 2015, SCEAPI, the self-developed interface for cloud service, has supported application platform construction on alpine environments, stem cells, high-energy physics and other research areas for 14 national and CAS key projects, including CAS alpine environment joint monitoring cloud, CAS stem cells and biomedical cloud, GridMol molecular visualization software, computational chemistry cloud service community, microgravity simulation platform, massively parallel hydrology platform, combustion kinetics center of Sichuan university, integration platform for materials science and so on. Relevant research won the best supercomputing application award of CAS in 2015. In addtion, it has provided computing resources for ATLAS - the world's top high-energy physics experiment in CERN, which got positive feedback from domain scientists.
CNIC provides the following software development service for customers.
1. Parallel Software Development and Performance optimization
2. Basic Parallel Software Development
Relevant services include: 1) providing a bundle of fundamental parallel utility software for scientific computing and data processing to facilitate users with their work; 2) self-developed fundamental parallel utility software; 3) importing open source projects and writing parallel software with independent intellectual property rights indicated by fundamental parallel modules.