This large tech museum in Padaborn Germany has SN2o18 Cray-2 on display.
Editor note: SN2018 is currently listed as having previously worked at National Energy Research Supercomputer Center (NERSC) @ Lawrence Berkeley Livermore Labratory and Internal machine at Cray Research, Inc (CRI). Not sure how it ended up in Europe unless it did some time at another undisclosed site. Was lent to museum by Cray Research Germany.
Museum link here
Heinz Nixdorf MuseumsForum
Phone +49 (0) 5251 306-600
The weather forecast with the help of the computer was at the beginning of electronic computing. John W. Mauchly wanted to use the enormous speed of the ENIAC at the time to numerically solve the extensive systems of equations that describe the weather. John von Neumann, who played a leading role in the further development of computers in the USA, then devoted himself to this task. “If it takes more than 24 hours to calculate the weather for tomorrow, we’ll drop it,” he reportedly said.
Fast computers were therefore required. At the beginning of the 1960s, Seymour Cray was able to demonstrate at Control Data with the CDC 1604 that a hundredfold increase in computing speed was possible with the available technology. A short time later he became self-employed and brought the legendary CRAY-1 onto the market. The supercomputer was born. Several processors simultaneously carried out the same calculation steps when solving systems of equations. Given suitable problems, this so-called vector computer could therefore achieve the multiple computing speed of a single-processor machine. The successor model, a 1986 Cray-2, is the main attraction of this exhibition area.
With the availability of fast microprocessors at the beginning of the 1990s, a variety of attempts were made to implement powerful multiprocessor systems. In addition to the family tree of vector computers, the exhibition shows the various structures or topologies in which multiprocessor systems were built.
Simulations are the field of application of the fast machines. In order to carry out a meaningful “virtual” car crash test, in addition to entering a wealth of design data and the parameters of the experiment, careful comparisons of the calculated data with the experiments actually carried out are required. As soon as the method delivers reasonable results, complex tests can be carried out as early as the design phase. The engineer can use it to test design ideas and their effects on the overall system in an early test phase. The simulation then replaces reality.
Today we all benefit from a significantly improved weather forecast. Today’s simulations on supercomputers can not only “calculate” the simulation once, but very often with different starting parameters, which shows how stable the results are.