Supercomputers are looking for an answer to the fundamental question of the universe

The international team of scientists reported in the online edition of Physical Review Letters on the landmark achievement in subatomic physics - Numerical simulation of kaon decay into two pions using the fastest supercomputers.

During the opening at the Brookhaven National Laboratory for this phenomenon, the authors were awarded the Nobel Prize in 1964. It was the first experimental evidence of parity violation charges - no symmetry between particles and antiparticles - and the first step to getting an answer to the question - why our world is made of matter rather than antimatter.

Calculations, modeling one aspect of this decay, required the development of innovative techniques and the use of the most powerful modern computers. They are based on new theoretical results obtained after 1964, and recent experiments performed at CERN, the European particle physics laboratory, and the Fermi National Accelerator Laboratory. The project required collaboration specialists Brookhaven Laboratory and Research Center of RIKEN, Columbia University, Connecticut, Southampton, Washington, and Edinburgh, the Max Planck Institute of Physics.

A special problem is the unprecedented scale of the scatter calculation is 18 orders of magnitude, from sub-atomic level interactions to the size of the bubble chamber. Calculations of the collapse, according to the methods of quantum chromodynamics (QCD) is happening at the nodes of the space-time lattice, performed the statistical algorithm of the Monte Carlo method. They demanded 54 million hours of supercomputer CPU IBM BlueGene / P, installed at the Argonne National Laboratory (ALCF). For preliminary calculations used a prototype computer QCDOC (QCD on a chip) and a series of IBM BlueGene.

This work has allowed experts to estimate the unknown quantity, is important for understanding kaon decay in the aspect of parity violation charges. Direct numerical calculation of this parameter, as well as the conversion is the result of increased accuracy would require more computer resources.

To do this, members of the international project are hoping to use the power BlueGene / Q, the next generation of supercomputers, IBM. These machines will have a speed of 10-20 times higher than that of modern analogs. Over the next few months, they plan to install many of the leading research centers, including the ALCF, Brookhaven Laboratory, University of Edinburgh and laboratory KEK (Japan).