The network's activities are organised into eight inter-related project areas which are addressed by the following Working Groups:
Person-in-charge: Tilo Wettig
Lattice computations break down into large tasks, requiring massively parallel "capability" computers, and smaller tasks needing several smaller "capacity" machines. CPU clock rates ceased to increase around 2004, due to the heat produced (the power wall). A low power/performance ratio is now the most important design parameter of supercomputers, with a move towards multicore CPUs. Two levels of parallelism, on- and off-chip, complicate the networking of such processors. Also, the communication bandwidth has to keep up with the increased on-chip compute-power.
The teams of Regensburg University, Wuppertal University and IBM, in collaboration with DESY, the Jülich Supercomputing Center and Ferrara University, have developed the QPACE supercomputer based on the IBM PowerXCell 8i (Enhanced Cell/B.E) processor. Universitá di Parma is part of a scientific collaboration aiming at defining architecture and software environment of a machine in the Pflops range. Eurotech collaborates with them on HPC development (Aurora supercomputer), based on Intel processors (Nehalem and follow-ups). The University of Edinburgh team collaborates with IBM Research USA in the design of a future BlueGene architecture. Members of the WG are discussing follow-up machines whose development will start in 2011. The final design choices have not yet been made.
Person-in-charge: Tony Kennedy
Lattice QCD codes that combine end-user-friendly, cross-platform compatibility with highly optimised, machine-specific assembler code for computationally-intensive core routines are needed. The free open source Chroma/QDP++/QMP package aims at this. New physics application software needs to be added to Chroma and, to exploit new computers and compilers, efficient assembler routines need to be written and existing low-level routines upgraded or migrated. In particular, the thread-level-parallelism of multicore CPUs needs to be optimally utilized.
Person-in-charge: Francesco Knechtli
Efficient numerical algorithms are essential to exploit new supercomputer infrastructure. In this area, a lot can be gained through increased cooperation between the applied mathematicians and lattice QCD algorithm specialists in our network.
Person-in-charge: Francesco Di Renzo
Physical processes can often be factorized into high momentum Wilson coefficients that are calculable in (lattice) perturbation theory ((L)PT) and non-perturbative matrix elements. LPT can be done in the traditional diagrammatic way or with the numerical stochastic method (NSPT). The two methods are often complementary.
Person-in-charge: Mike Peardon
The hadron spectrum is one of the key tests of our understanding of non-perturbative aspects of QCD, and Lattice QCD predicts states (e.g. glueballs, quark-gluon hybrid states, tetraquarks, excited baryons etc.) which are being sought experimentally. Recent advances in all-to-all propagator techniques allow us to incorporate the previously difficult flavour-singlet sector with high precision, extend our operator basis to include multi-quark states and start to explore strong decay channels and scattering states.
Person-in-charge: Constantia Alexandrou
Hadrons and particularly the nucleon are the basic constituents of visible matter in the universe and their structure is intensively studied experimentally. It is vital that these programmes are supported by theoretical calculations of the nucleon's unpolarised, polarised and transversity structure functions, disentangling valence, sea quark and gluon distributions, as well as by calculations of generalised parton distributions (GPDs) and distribution amplitudes. QCD calculations of EM transition rates will help to pin down the nature of narrow charmonium resonances. EM transition form factors between baryons are of major experimental interest as well.
QCD weak matrix elements are needed to relate experiment to Standard Model parameters and, eventually, to identify new Beyond the Standard Model (BSM) physics. Accurate theoretical predictions for the D meson sector are needed (as evidenced by the very recent discovery of CP violating D-decays).
Person-in-charge: Edwin Laermann
At high temperatures and/or high baryon densities, QCD undergoes a transition from the hadronic phase to the quark-gluon plasma (QGP) phase. The high-temperature transition took place during the evolution of the early universe, while the high-density phase(s) may be realized at the cores of supermassive stars. A major goal of the experimental programmes at RHIC, LHC and FAIR is to study the properties of these transitions and of the QGP.
Proper interpretation of the experimental results requires a good understanding of the thermodynamics of strongly interacting matter. Extrapolating results to conditions not easily reached in the laboratory makes theoretical studies essential. Only by an interplay between theory and experiment can we obtain all the information needed e.g. for certain dark matter relic density computations in cosmology and the analysis of superdense systems in astrophysics.
Person-in-charge: Carlos Pena
The spontaneous breaking of chiral symmetry and quark confinement are the two fundamental non-perturbative features of QCD. The former is almost entirely responsible for the masses of light hadrons while heavy quarkonia can be understood in terms of a confining potential. These two properties of the QCD vacuum are preserved in the large-N limit of QCD. In SUSY QCD, the large-N limit at large 't Hooft coupling is thought to be connected to weakly coupled supergravity near the surface of five dimensional anti-de Sitter space (AdS-CFT correspondence). It is interesting to check in how far QCD inherits features of its SUSY QCD sister theory, and to compare the spectrum with classical SUGRA predictions/expectations. It is another exciting theoretical challenge to find discretizations in which a SUSY-respecting continuum limit can be taken, allowing us to go beyond QCD and to simulate SUSY QCD on a computer. The LHC might reveal a composite Higgs-particle, bound by a new Yang-Mills theory similar to QCD (technicolor). Different N, fermion numbers and representations can be explored, enabling us to constrain the space of candidate theories.