QUENCHED COMPUTATION OF THE LIGHT HADRON MASS SPECTRUM
Ever since the pioneering numerical simulations of lattice QCD in 1981, the calculation of the light hadron spectrum has been a fundamental subject in lattice QCD. QCD simulations on the lattice, however, require a huge amount of computer time. Therefore, most large scale simulations have been performed using an approximation of neglecting the effects of quark paircreations and annihilations in the vacuum (quenched approximation). This reduces the computer time by a factor more than 100 and enables QCD simulations on relatively large lattices with high statistics. It is expected that the deviation in the light hadron spectrum due to the quenched approximation is about 10%. A longstanding issue in lattice QCD has been whether the quenched QCD reproduces the experimentally observed properties of hadrons at the level of 10%, and, furthermore, whether deviations due to the quenched approximation can be established. Solving these problems require a precision much better than 10%. This has proven quite difficult even with the powerful supercomputers of early 90's. A breakthrough was achieved by the CPPACS Collaboration by exploiting the computing power of the CPPACS computer. In the figure ``Development of the computer power devoted to lattice QCD calculations'' [JPEG  GIF  PDF], we show an estimate of the number of floatingpoint operations devoted to quenched hadron spectrum computations by major research groups. The CPPACS enabled a calculation which is one to two orders of magnitude larger than previous studies. We have carried out quenched simulations of the light hadron mass spectrum with a larger latticesize and higher statistics than previous studies, and have succeed in reducing the errors for light hadron masses down to about 13%. This has enabled us to reach precise and clear conclusions about the effects of the quenched approximation in physical observables. The results for light hadron mass spectrum are summarized in a figure ``Hadron mass spectrum from quarks and gluons'' [JPEG  GIF  PDF], where experimental values as well as the results of the best previous study by the GF11 Collaboration are also plotted for comparison. We found that
We also studied other physical observables such as hadron decay constants and quark masses. In a figure ``Determination of fundamental parameters in Quantum Chromodynamics'' [JPEG  GIF  PDF], results for quark masses obtained in recent largescale simulations are summarized. The CPPACS drastically increased the quality of predictions from QCD. 
FULL QCD SIMULATIONS
Having observed an unambiguous deviation of the quenched light hadron spectrum from the experiment, we have started, as a logical next step, a systematic simulation of full QCD, i.e., QCD taking into account the effects of pair creations/annihilations of quarks in the vacuum (dynamical ``sea'' quarks). Major sea quark effects are expected to be coming from the lightest two quarks, ``up'' and ``down''. Therefore, we are carrying out simulations with two flavors of sea quarks, to be identified with dynamical up and down quarks. Preliminary results for light hadron spectrum have been presented at international conferences (see the publication list below). They indicate that the discrepancy with experiment is much smaller than that obtained for quenched QCD. The remaining small discrepancy might be due to the quenched treatment of heavier quarks, especially the third quark ``strange''. Studies of finitetemperature QCD relevant to the evolution of the early Universe, as well as studies of the nature of heavy hadrons made of heavy quarks ``charm'' and ``bottom'', are also well underway. 
LIST OF PUBLICATIONS

Back to
