In the current industrialized world with increasing technological advance requirements, the need for developing tools has tremendously increased to achieve the desired goals. One important tool is to build up high speed computational devices with high memory capacities. Indeed, to solve the problems in simulation, design and other related areas, all industries and universities require to fulfill very time-consuming computer calculations. Evidently, large time consuming jobs cannot be readily handled with ordinary computers. In other words, the computational requirements for solving problems and designing real systems are beyond the capabilities of conventional computers which are available now.
The need for increasing the accuracy in simulations and capturing the real phenomena would normally increase the amount of calculation. For example, in the numerical solution of fluid flow using the Fluent software, using a high number of elements for a complex geometry with complicated fluid flow requires very long time to be analyzed by an ordinary computer, such that in some cases users have to ignore the details analysis of the problem.
With increase in the amount of calculations, more attention should be dedicated to computer facilities to reduce the time needed for analyzing the industrial problems and to speed up the related calculations. The limitations in accessing to supercomputers in our country, promotes the idea of parallel processing which not only resolves the problem but also enables us to compete with other international institutions. Using this idea, it is possible to achieve supercomputers speed using ordinary servers. This idea has been common used in building many of current supercomputers, such that these supercomputers are built from ordinary computers. The literature shows that such supercomputers are found among the world top 10 supercomputers.
Essentially, the need for increasing numerical simulation in various fields of science and technology necessitates more computational capacity demand than the past. Due to importance of numerical simulation of aerospace engineering systems, the Aerospace Engineering Department of Sharif University of Technology established a high performance computing laboratory to improve its research computational capabilities. The laboratory was established under supervision of Prof. Darbandi and a number of his research assistants including Ghafourizadeh, Abrar, Setayeshgar, Mehraban, benefiting from the financial support received from the Department of Aerospace Engineering in autumn of 2009. Some technical information about the operating cluster is presented next. An Overview of the assembled cluster is observed in the following picture.
So far many applications and softwares have been installed and are under service. The softwares cover various areas such as industrial fluid flow simulation, finite element analysis, CFD, computational grid generation, numerical analysis and control, MPI libraries, different languages compilers and so on.
Copyright 2010 Aerocluster.sharif.edu, All rights reserved, Designed by: Javad Ghafourizadeh