Persian
AeroSpace Dep Logo HPCL
Sharif Uni Logo
Aerospace Engineering Department Sharif University Of Technology
Welcome to high performance computing lab website.

Design and set up of Aerospace Engineering Department PC-Cluster

The system was first designed and established under the supervision of Prof. Darbandi and his research assistant Mehdi Bostandoostnik in 2001. The old system had a total computational capacity of 44 gigaflops consisted of 9 PCs having the specifications of 2.4 GHz single-core processor, 256 MB RAM and 60 GB HDD. The nodes were interconnected with a 100 Mbps hub. Many packages and compilers had been installed on the system, e.g. Fortran, C, C++ with Linux RedHat 7.3 as the Operation System and the data had been exchanged among PCs by MPI library. The math libraries such as PETSc, and Hypre had been installed and used in parallel programmings. The system was upgraded in 2006 by Majid Ghafourizadeh. In this regard, the communication switch was upgraded to 1000 Mbps and the Operating system to Linux RedHat 9. All of the software were upgraded as well. In 2007, Majid Ghafourizadeh and Alireza Setayeshgar tested and improved the parallel programming to Windows platform. Additionally, the diskless system was tested and performed as well.

Design and set up of Agricultural Engineering Research Institute PC-Cluster

The system was designed and established under the supervision of Prof.Darbandi and the his research assistants Majid Ghafourizadeh and Alireza Setayeshgar in 2008. The system had a total computational capacity of 173 gigaflops and was consisting of 9 PCs with an individual specification of 2.4 GHz quad-core processor, 4 GB RAM and 250 GB HDD. The nodes were interconnected using a 1000 Mbps switch. To monitor the computational nodes, we used a KVM switch. The picture of this cluster is observed here.



Design and set up of Aerospace Engineering Department Cluster (Aero-Cluster)

The system was designed and established under the supervision of Prof. Darbandi and his research assistants Majid Ghafourizadeh, Bagher Abrar, Alireza Setayeshgar and Yones Mehraban in 2009. The system had a total computational capacity of 1445 gigaflops. It was consisted of a server with specifications of 3.0 GHz octa-core processor, 16 GB RAM , 1500 GB HDD as manager node and 16 servers with specifications of 3.0 GHz octa-core processor, 8 GB RAM , 1000 GB HDD as computational nodes. The nodes are interconnected with a 1000 Mbps switch. The system has also been equipped by a KVM switch to enable computational nodes monitoring and a temperature warning system. Simulation softwares such as Fluent, Ansys, MATLAB and libraries for parallel processing have been installed and its parallel performance was evaluated. A picture of this cluster is observed here.



Copyright 2010 Aerocluster.sharif.edu, All rights reserved, Designed by: Javad Ghafourizadeh