High performance computing

from Wikipedia, the free encyclopedia
Interview between Holger Klein and Bernd Mohr from the Jülich Supercomputing Center on the subject of high-performance computing.

High Performance Computing ( English high-performance computing , HPC ) is an area of the computer-based computing. It includes all computing work that requires a high level of computing power or storage capacity.

High-performance computers are computer systems that are suitable for processing high-performance computing tasks.


High performance computing

Due to the rapid development of computing technology, there are no permanent formal criteria for the exact delimitation of high-performance computing from the other areas of computer-aided computing. However, it is generally recognized that computing applications whose complexity or scope makes computation on simple workstation computers impossible or at least nonsensical fall into the field of high-performance computing.

High-performance computing is made possible in the first place by the architecture of high-performance computers, which is geared towards parallel processing. In order for applications to be able to use this architecture, they must be programmed in parallel . This can be done with OpenMP and MPI , for example.

High performance computer

Similar to high-performance computing, the term high-performance computer can hardly be formally defined. An essential feature, however, is an architecture geared towards parallel processing.

One of the high-performance computers is, on the one hand, highly parallel supercomputers , and, on the other, computer clusters , which can be both distributed ( distributed computing , grid computing ) and organized in a local network. Other computer systems with a special architecture, for example with a very large, efficiently organized memory, can also be referred to as high-performance computers.

At present, computers in the teraflops performance class and above are particularly high-performance computers.


In scientific computing in particular, high-performance computing is increasingly important as a tool for calculating, modeling and simulating complex systems and for processing huge amounts of measurement data. Such applications can be found today in practically all areas of natural and technical sciences; Typical areas of application are meteorology and climatology , astro and particle physics , systems biology , genetics , quantum chemistry and fluid mechanics .

There are also high-performance computing applications in commercial computing. Many of them are of scientific origin (e.g. weather forecast, crash test simulation, flow simulation in aircraft construction), but there are also applications without a scientific character, e.g. B. in the generation of animated films.

Occasionally, applications for high-performance computing are also counted which, as individual application instances, do not make any unusual resource requirements, but which typically run in very large numbers at the same time and develop a complexity from this; Typical examples can be found in database applications in the financial sector and in the switching of telephone connections.

See also

Web links

  • Top 500 - List of the 500 fastest computers

Individual evidence

  1. ^ Resonator podcast of the Helmholtz Association : High Performance Computing (episode 61, June 5, 2015)