Along with the development of economic theories, quantitative analysis has established itself as a major research method in the field of economic and financial studies and has called on the need of proving high-performance computer software and hardware solutions. Firstly, the high-performance computation environment can fully enhance the efficiency of data analysis in scientific researches and greatly improve the pattern of economic research and education. Nowadays, high-performance computation has played an important role in macroeconomics, microeconomics, experimental economics and econometrics. Especially in the field of modern financial studies, which usually crossovers with mathematics, statistics and econometrics, the high-performance computer has provided a solid foundation for the further development of financial econometric modeling and computation. Secondly, because of the unrepeatability of economic phenomena, Monte Carlo method, which is strongly based on the large-scale simulation computation, has become a widely used method in modern economic studies. Furthermore, the verification of newly developed econometric model also requires the participation of high-performance computation.
After half a year’s lengthy design and planning, WISE partnered with Microsoft, Dell and SAS, has set up a national leading high-performance computer cluster (HPCC) at Xiamen University (XMU) on April 2007, which is a hi-tech achievement for “985project (Ⅱ)”,and therefore has filled the vacancy of economic simulation computation in China.
The technology of high-performance computation aims to achieve increased performance by splitting the computational tasks across many different processors or computer clusters. Based on its systematic setting of resources, the technology provides the possibilities for parallel computation, transmission and storage, and is most commonly used in scientific computing, which is relative to large-scale parallel computing resources, real-time analysis and control, etc. On recent stage, the usage of high-performance computation has been considered to be a criterion of academic and research capacity.
With regard to the hardware of the newly set up HPCC at WISE, we has chose the latest High Performance Computer Cluster from DELL and has constructed 25 high-performance parallel computing nodes at the first phase. By the equipment of 3.0G*100 XEON CUP, 100G RAM and 10BT storage, the HPCC can achieve a computing capacity of 512G flops theoretically and of 250 billion times per second using gigabit Ethernet in reality, which is enough for fully operating 100 parallel computing tasks. At the second phase, we are planning to construct 75 more nodes, which will provide a computing capacity of 10 trillion times per second or more based on 400 pieces of 3.0G XEON CPU.
As to the operation system, the Microsoft Compute Cluster Server 2003 (CCS) has successfully brought the highly complex computation application together with the familiar Windows operating system and hence provides a security-enhanced, maneuverable, high-performance parallel computing solution, which can benefit the agents with the efficiency of searching for the key issues.
For the software application, the HPCC has integrated with SAS X64 Sever Edition. With the help of MP CONNECT, parallel processing is enabled and the idle processor can be used to reduce the time required to analyze huge amounts of data. Besides, installations of MATLAB, GAUSS, R and STATA are also on the schedule.
Since the HPCC came into full production in April, the empirical and simulative research environment at WISE has significantly improved and a series of studies has been promoted, which includes the empirical study of anti-money laundering partnered with SAS and the People’s Bank of China Xiamen Branch, the collaborative research of risk forecasting and management with the Econometrics Studies Center of Shanghai Academy of Social Sciences, the collaborative research on the high-frequency data in Chinese stock market with Singapore Management University and the collaborative research on the volatility of stock markets with Cornell University and University of North Carolina at Charlotte.
“Cooperating with WISE to develop the HPCC is a great experience. We will continuously focus on the application of high-performance computation in the field of economic studies and fulfill the goal of providing a familiar and maneuverable Windows environment to the economists for model development and data analysis.”
-------Timothy Chen, Microsoft Corporate Vice-President and CEO, Greater China Region
“We happily work with WISE in this collaborative project. Dell’s 9th generation severs is specially designed for the high-performance computation and Microsoft CCS. We hope this generation of the products can provide a solid foundation for the Chinese further development in the field of economic simulation computation.”
-------David D. Miller, President of Dell China
“SAS and WISE have successfully build up a solid cooperative relationship and have jointly set up the SAS Center of Excellence in Econometric at XMU. This cooperative project will further utilize the potential of SAS in the field of huge amounts data process and intelligent data mining. We will continuously hunt for the opportunities of cooperation with economic institutes and contribute to the development of Chinese academic research.”
--------Alex Wong, President for SAS Greater China
“Joyfully note that WISE has acquired a national leading high-performance computer cluster. This installment of HPCC will definitely contribute to the solution of more and more complex economic problem. We want to share our experience in the field of engineering parallel computation with WISE and jointly enhance the application of high-performance computation in Chinese economy.”
-------Simon J. Cox, Director of Microsoft High Performance Center, University of Southampton, UK
(WISE)
