Supercomputer History

1247 Words3 Pages

1 .Introduction
The purpose of this report is to study the current state of supercomputers and to briefly discuss predicted advancements in the field..A brief history about the supercomputer and its operation is initially outlined.
Supercomputers are the supreme machines in the computer world. Supercomputers includes thousands of times the computing power of a desktop and cost several millions of dollars. They fill large halls, which are chilled to prevent their thousands of microprocessor cores from overheating. And they perform billions, or even thousands of billions, of calculations per second.
All of that power means supercomputers are better for tackling any big scientific problems, from uncovering the origins of the alien planets to delving into the patterns of protein folding that make life possible.
So in this report the reader can get a good knowledge about supercomputer.
2. What is a Supercomputer?
A supercomputer is a very powerful digital computational machine capable of processing trillions of commands per second. It is actually at the leading edge of current processing capacity, especially its unbelievable speed of calculation. They are extremely useful for doing analysis of overtly complicated problem sets related to industry as well as government. So they are usually owned by countries or corporation and never for personal use.
A supercomputer can be defined as a powerful mainframe computer .The chief difference between a supercomputer and a mainframe computer is that a supercomputer channels all its power into executing a few programs as fast as possible, where as a mainframe computers uses its power to execute many programs concurrently..
Super computers are characterized by very high computational speeds an...

... middle of paper ...

...all the multi-threading math libraries and compilers to build your parallel computing programs.
8. Network the compute nodes all together. The head node sends the tasks to the other compute nodes .Use a private internal network to connect all the nodes in the cluster.
The head node work as a NFS, PXE, DHCP, TFTP, and NTP server over the Ethernet network.
We must separate this network from other public networks, which ensures that broadcast packets don't interfere with other networks in your local area network.
9. Test the cluster. The final thing we may want to do before releasing all this mashine power to your users is test it's performance. The HPL (High Performance Lynpack) is a famous choice for measuring the computational speed of the cluster computer. we need to compile it from source with all optimizations our compiler offers for the architecture we chose.

Open Document