Search

link to homepage

Navigation and service


Supercomputers and Simulations

Simulations on supercomputers of the highest performance class are an indispensable tool for science. For researchers they are a bridge between theory and experiment, enabling them to obtain insights and knowledge which would otherwise be impossible for physical, technical, financial, or ethical reasons.

Scientists create virtual universes on their computers and use them to explore the basic conditions for the origin of life on Earth. Supercomputers can radically reduce prototype cycles in mechanical engineering and aircraft construction and also in the pharmaceutical industry. A particular challenge is the simulation of the human brain. This is not only relevant for the diagnosis and treatment of brain diseases but could also provide new impetus for electronic data processing since, in its own way, the human brain is a high-performance and energy-saving "supercomputer" from which information scientists have a lot to learn. Thus, supercomputing is itself a branch of science, while simulation science is a key technology for other disciplines.

Simulating Complex Effects

Simulations are indispensable for investigating complex physical and mechanical effects in the nanometre or quantum range which could be exploited for novel technologies in data processing and storage. Conventional technologies are rapidly reaching their limits. Computers have to work ever faster and store more data while at the same time consuming less energy.

This is why scientists are working on concepts that do not use the electric charge of electrons for data processing but rather their magnetic moment – their "spin". Electrons will then no longer have to flow, but must merely transfer the spin information.

In 2013, by making use of simulations on supercomputers, an international team of physicists discovered a promising new physical effect – the spin Cherenkov effect. This is a magnetic version of the well-known Cherenkov effect, which occurs when charged particles travel through water faster than the speed of light. The simulated effect would enable spin waves to be generated more easily in magnetic materials with a defined frequency than had previously been assumed. If this effect could be harnessed then appropriately designed computers could work faster and at the same time more energy-efficiently.

On the Road to Quantum Computing

Researchers hope to use quantum computers to solve a number of problems that remain as yet insoluble. Whereas classical computers store information in the form of bits, quantum computers work with quantum bits. These qubits are based on the quantum mechanical properties of the electron, which is not regarded as a particle but as a wave. In this way, various quantum mechanical states can be superimposed on each other, which may lead to enormously increased speed. Theoretically, numerous algorithms could run in parallel on such a system. And it also facilitates a logic that does not only use zero and one as values but also intermediate states. Physicists at Jülich's Peter Grünberg Institute are seeking ways to realize quantum information processing. There is still a long way to go before such a machine becomes a practical proposition. However, scientists at the Jülich Supercomputing Centre are already using simulations to determine in detail how conventional applications, such as Shor's algorithm for factoring large numbers, can be implemented on quantum computer systems.

Back in 2010, the Jülich supercomputer JUGENE simulated the world's largest quantum computer system at that time with 42 bits.

Further information

more on Supercomputing
more on Simulation Science
Jülich Supercomputing Centre (JSC)
Peter Grünberg Institute - Theoretical Nanoelectronics (PGI-2 / IAS-3)
Press Release: "Electrically Controllable Qubits Made a Reality"
Press Release "Magnetic Version of Cherenkov Effect Discovered"


Servicemeu

Homepage