Return to search

Taking lattice QCD beyond the quenched approximation

This thesis is mainly concerned with the problem of generating gauge configurations for use in Monte-Carlo lattice QCD calculations that include the effects of dynamical fermions. Although algorithms to do this have been in existence for some time the computing power necessary for their application at a scale where physically relevant results can be obtained has only recently become available, so these large scale dynamical simulations are still a new feature of lattice QCD. The emphasis here is on the new experiences gained from the design, implementation and development of a particular dynamical gauge configuration algorithm, and from its initial use in production. The intention is that this will facilitate future computations where the effects of dynamical fermions in QCD can be systematically explored, and the further development of better algorithmic techniques. The first chapter outlines those features of lattice gauge theory computations that are salient to this work, concentrating particularly on the motivations for and consequences of going beyond the quenched approximation and on the properties of Markov processes used in the generation of gauge configurations. The second chapter introduces the main types of algorithm employed for dynamical gauge configuration production, viz. the multiboson algorithm and Hybrid Monte-Carlo, and explains how they work. In chapter 3 the implementation of the chosen algorithm (Generalised Hybrid Monte-Carlo) is described, along with various algorithmic investigations, coding developments, performance evaluation, and a description of the procedures used in the verification of the code. Finally, some results are presented from the first large scale production runs on the Cray-TBE which attempt to put the algorithm work into a physical context.
Date January 1998
CreatorsSroczynski, Zbigniew
PublisherUniversity of Edinburgh
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation

Page generated in 0.0025 seconds