Scientific computing > Supercomputers and grids
 
Tehdyt toimenpiteet

CSC's Server Environment

Mathematical modeling, computer-aided simulations, data analysis, and the use of databases form a part of modern science. The versatile computer environment enables large-scale simulation of research phenomena. Researchers have remote acccess to CSC's computers and scientific software. This ensures that scientific research is equally possible in different parts of the country. Finnish researchers can connect to CSC's computing environment via the Funet network. Several services, e.g., some of the databases, are available directly via Web browser.

Massive parallel computing

CSC's supercomputers are dedicated to run powerful parallel applications (such as simulations of materials physics). With parallel computing, the computing capacity can be increased a 100-fold compared with the computing done on a single processor. Parallel computing reduces the time needed to solve a complicated problem from several CPU years to few weeks. CSC's most powerful computer Cray XT is capable of computing 10,000 billion floating point operations per second. It equals to 1800 Pentium P4 processors.

Massive memory

The massive memory of supercomputers benefits particularly researches who create models of physical phenomena involving massive data and large databases. Powerful numerical simulations give insight, e.g., into the behavior of astrophysical bodies, new molecules and their properties, structural engineering, and problems related to computational fluid dynamics. Supercomputers also enable investigations of whole ecosystems or global environmental models.

Archiving data

Scientific computing involves compilation of large amounts of data, and tape technology is needed to save data in a cost-effective and secure manner. Researchers can archive the data they need for their computing.  A new application area for the server is archiving Finnish www pages in collaboration with the National Library of Finland. CSC provides three petabytes of data storing capacity. It corresponds to data on 23 billion pages in a book. If the books were placed on a bookshelf the shelf has to be a 1,000 kilometers long. A petabyte corresponds to 1,540,000 CDs. If one CD is about one mm thick, this would amount to a pile more than 1.5 kilometers high.