Computational Theoretical Physics Laboratory (CTP Lab) aims to reveal the complicated aspects of physics quantitatively, that are not calculable by just pencils, papers and brains, utilizing high-performance parallel computers.
Recent rapid development of computers has altered the landscape of computational science and enabled us to tackle the area of researches that was beyond reach a decade ago. We use two newly-introduced supercomputers to perform the present research subjects in much larger scale, and also to develop the new research areas that are now within reach.
The two supercomputer systems are equipped in the institute; one is dedicated to the studies of the theoretical elementary particle physics and hadron physics (named "Phi"), and the other is to the astrophysics simulations.
The Phi system has a feature that it consists of, in addition to the ordinary CPUs, GPUs (Graphics Processing Unit) that are applied to general purpose programming to accelerate the computation.
Combining CPUs and GPUs the whole system attains 62TFlops (62 trillion floating point operations per second) in its peak performance, which equals to one of top 20 supercomputers in Japan (in Oct. 2010).
The system for the astrophysics has high-speed interconnect between computer nodes and large-volume storage units that are required for astrophysical simulations.
CTP Lab conducts extensive numerical studies of the elementary particle physics and astrophysics using these high-performance computer systems and developments of algorithms and program codes to perform these studies more efficiently.
The major research topics for elementary particle physics utilizing the Phi system are to carry out numerical simulations of quantum field theory for the physics beyond the standard model and to solve the gauge dynamics beyond the reach of conventional numerical simulations.
The Standard Model is the theory of elementary particles that involves three of the four fundamental interactions of nature (strong, weak, and electromagnetic forces, except for gravitational force) in a unified way. Still all of these forces have not been fully understood. It is known that, in the framework of the Standard Model, the quarks and leptons acquire their masses through the spontaneous breaking of the symmetry of the electromagnetic and weak forces. But the mechanism of this symmetry breaking is still unclear. It is expected that the LHC experiments running at CERN will provide valuable inputs to this puzzle.
CTP Lab is studying to clarify this unknown mechanism and to explore theories beyond the standard model in cooperation with Division of Theoretical Particle Physics. One such theory of our concern is called the technicolor model in which the electroweak symmetry breaking results through the dynamics of strongly-coupled systems. These theories are much constrained by several conditions (e.g. the dependence of the coupling constant on the energy scale is very weak, and the theory behaves strongly-coupled in the low-energy region). To study whether or not there are any theories that satisfy these conditions, various strongly-coupled models have to be examined elaborately. Because the perturbative analysis is not applicable there, some nonperturbative approach is indispensable. One of the possible tools for the non-perturbative calculation is the numerical simulation with computers. The researches with intensive use of the numerical facilities may clarify whether there exists any theory that fulfills requirements of the technicolor model and, if any, what sort of features it exhibits.
Another topic of our research with the Phi computer is the nonperturbative study of the quantum chromodynamics (QCD) that explains the strong force.
QCD is also strongly-coupled theory, and therefore the numerical investigation of the lattice QCD is the powerful tool for the verification of the standard model including QCD, direct calculation of nuclei from QCD, and also the analysis of its characteristics.
The phase structure of QCD and the nature of hadronic phase at high temperature and high density region are also studied which are directly related to the quark-gluon plasma (QGP) phase found in the recent experiments in Brookhaven National Laboratory.
Furthermore numerical approach is applied to perturbative calculation of the quantum electrodynamics (QED) concerning the ultra high precision calculation of the anomalous magnetic moment of leptons that leads to the most stringent test of QED and to the exploration of new physics.
The system for the astrophysical simulations is applied to the researches concerning the nature of dark energy, neutrino masses, non-gaussianity of the density distribution in the early universe, and other topics.
The dark energy and dark matter, whose existence has been confirmed through the remarkable development of astronomical observation techniques, are still obscure of their origin or nature. Clarifying their nature is one of the most important issues to be solved both in the fundamental physics and astronomy. Since the dark energy and dark matter occupy 96% of the total energy of the universe, they are considered to govern the history of the universe and the evolutions of the galaxies, clusters of galaxies, and the large-scale structure of the universe. Therefore, it is expected that the abundance and the nature of these dark sides will be strongly constrained through the detailed examination of the data on the time-evolution of the galaxy distribution and the weak lensing effect for the galaxies that are obtained by large-scale galaxy survey.
Toward these aims CTP Lab takes part in the Weak Lensing Survey Project (HSC Survey) that makes use of wide viewing angles of Subaru telescope and provides theoretical estimates of observation designs, simulation data, and developments of theories. Observation will yield much more precise data in the near future. To confront these results the theoretical result has to be precise enough. The evolution of the large-scale structure of the universe via gravitation is essentially non-linear phenomena, so that extensive numerical studies are required. Therefore using the computer system for the astrophysical simulations we are working on the clarification of the features of the time-evolution and dispersion for gravitational lensing effect, the elaboration of the likelihood function, and so forth.
