Chulwoo Jung

HARDWARE AND SOFTWARE STATUS OF QCDOC.

QCDOC is a massively parallel supercomputer whose processing nodes are based on an application-specific integrated circuit (ASIC). This ASIC was custom-designed so that crucial lattice QCD kernels achieve an overall sustained performance of 50% on machines with several 10,000 nodes. This strong scalability, together with low power consumption and a price/performance ratio of

Status of the 0.8 teraflops supercomputer at Columbia

1 per sustained MFlops, enable QCDOC to attack the most demanding lattice QCD problems. The first ASICs became available in June of 2003, and the testing performed so far has shown all systems functioning according to specification. We review the hardware and software status of QCDOC and present performance figures obtained in real hardware as well as in simulation.

Status of and performance estimates for QCDOC

The first stage in the construction of the 0.8 Teraflops Supercomputer at Columbia, a working, two node parallel computer, has been successfully completed. The next stage, a 512 node, 26 Gigaflops prototype, is in its final construction phase. A general description and current status of the hardware and software is presented.

Status of the QCDSP project

Abstract QCDOC is a supercomputer designed for high scalability at a low cost per node. We discuss the status of the project and provide performance estimates for large machines obtained from cycle accurate simulation of the QCDOC ASIC.

QCDSP — A status report

Abstract We describe the completed 8,192-node, 0.4Tflops machine at Columbia as well as the 12,288-node, 0.6Tflops machine assembled at the RIKEN Brookhaven Research Center. Present performance as well as our experience in commissioning these large machines is presented. We outline our on-going physics program and explain how the configuration of the machine is varied to support a wide range of lattice QCD problems, requiring a variety of machine sizes. Finally a brief discussion is given of future prospects for large-scale lattice QCD machines.

The QCDSP project —a status report

Abstract The QCDSP machine at Columbia University has grown to 2,048 nodes achieving a peak speed of 100 Gigaflops. Software for quenched and Hybrid Monte Carlo (HMC) evolution scheme has been developed for staggered fermions, with support for Wilson and clover fermions under development. We provide an overview of the runtime environment, the current status of the QCDSP construction program and preliminary results not presented elsewhere in these proceedings.

Photon structure functions from lattice QCD

We give a brief overview of the massively parallel computer project underway for nearly the past four years, centered at Columbia University. A 6 Gflops and a 50 Gflops machine are presently being debugged for installation at OSU and SCRI respectively, while a 0.4 Tflops machine is under construction for Columbia and a 0.6 Tflops machine is planned for the new RIKEN Brookhaven Research Center.

Monopole-antimonopole condensation in the interpolating Georgi-Glashow model

Abstract We calculate the first moment of the photon structure function, 〈x〉− = ∫01 dx, Q2), on the lattice using the formalism developed by the authors. Different lattice spacings and volumes are studied to estimate the systematic errors present in lattice simulations. Also, theNf = 2 dynamical configurations, generated by SESAM collaboration, are studied for the effect of quark loops on (x)λ.

Relating U(1) monopole configurations to SU(2) saddle-point configurations

Abstract We study the three dimensional Georgi-Glashow model (which interpolates smoothly between pure U(1) and SU(2) limits) using a constrained cooling which preserves t Hooft-Polyakov monopoles. We find that the monopole-antimonopole condensation gives an area law for the Wilson loops. The monopole contribution to the string tension is close to the Monte Carlo value in the intermediate region.

Analysis of saddle-point configurations in 3-dimensional SU(2) lattice gauge theory

Abstract We have studied field configurations of the 3-dimensional Georgi-Glashow model which interpolate between the U (1) and the SU (2) limits. In the intermediate region, these configurations contain t-Hooft-Polyakov monopoles. We use cooling and extremization to find these configurations and investigate their evolution as we adiabatically move towards the U (1) and the SU (2) limits. We also evolve an SU (2) saddle point solution towards the U (1) limit to see the relation between the unstable solutions in the SU (2) theory and the stable ones in the U (1) theory.