D. Olson, Lawrence Berkeley National Laboratory
J. Yang, University of California, Los Angeles
P. Nevski, Brookhaven National Laboratory and the STAR Collaboration

Research Objectives

The STAR detector at Brookhaven National Laboratory (BNL) is designed to study the collision of heavy nuclei at very high energy. Its goal is to investigate nuclear matter at extreme energy density and to search for evidence of the phase transition between hadronic matter and the deconfined quark-gluon plasma. STAR and the RHIC accelerator (Relativistic Heavy Ion Collider) will generate over 300 terabytes of data each year, with an additional large volume of simulated data needed for data analysis and for understanding the performance of the detector. The generation and analysis of this simulated data is the objective of this research.

Computational Approach

	This simulation shows a single event, the collision of two gold ions with a center-of-mass energy of 200 AGeV. The color code indicates hits in the various subdetector components as well as indicating the momentum of particles. The image on the top is a perspective view near one end of the detector, looking roughly along the beam axis. The bottom image is a side view of the same event.

Physics results from experimental relativistic heavy ion collisions are derived from statistical analysis of large numbers of events (collisions of individual atomic nuclei). The theoretical models are implemented as Monte Carlo codes that describe the final state of each of the thousands of particles that are produced in these collisions. We use a number of these theoretical codes (VENUS, HIJING, RQMD, and others) to produce large samples of events. A simulation code called GEANT is used to propagate each of these thousands of particles through the material of the STAR detector and compute the reactions and energy deposition that occurs throughout the detector. These theoretical model codes and the detector simulation code are run on the MPP system utilizing the natural parallelism of the problem, namely that each event is independent, so that different events are computed in parallel on the various processor nodes.

Accomplishments

Over the past two years, STAR has generated a large set of simulated data on the Cray T3E. Approximately 250K PE-hours were used in fiscal 1999 and over 6 terabytes of simulated data produced. These data have been invaluable for understanding the detector response of STAR and developing analysis algorithms. They were essential as input for two large-scale Mock Data Challenges (MDC) at the RHIC Computing Facility at BNL, where the STAR primary data will be stored and first analyzed. Mechanisms were developed to efficiently transport large volumes of STAR data over the network between computing facilities spread across the country, a capability that will be crucial for the distribution of real STAR data. As a result of these efforts, STAR is now confident that the first data can be reliably handled and efficiently processed to extract the physics.

Significance

Publications

J. W. Harris et al. (STAR Collaboration), "STAR physics in the first two years at RHIC," Proc. 15th Winter Workshop on Nuclear Dynamics, January 1999, Park City, UT, Advances in Nuclear Dynamics, Plenum Press (in press).

T. J. Hallman et al. (STAR Collaboration), "The STAR scientific program," Proc. XIV Int. Symp. on High Energy Physics Problems, Relativistic Nuclear Physics, and Quantum Chromodynamics, August 1998, Dubna, Russia (in press).

H. Caines et al. (STAR Collaboration), "The year-one physics capabilities of STAR," Proc. APS Centennial Meeting, March 1999, Atlanta, GA, Relativistic Heavy Ion Mini-Symposium, World Scientific (in press).

http://www.star.bnl.gov/