|
|
|
|
Stimulating Relativistic Heavy Ion Collision Events in the STAR Detector
November, 1998
The Relativistic Heavy Ion Collider (RHIC) at Brookhaven National Laboratory (BNL) will collide beams of nuclei (as light as protons and as heavy as gold) at energies of up to 200 GeV per nucleon. At these energies, it should be possible to produce (for a very short time) a form of matter predicted by the Standard Model of theoretical physics, but never yet observed experimentally. In this state, quarks and gluons would not be bound into the ordinary nuclear matter we are made of and which surrounds us. Such "Quark-Gluon Plasma" (QGP) may have existed in our Universe immediately after the "Big Bang", and may still exist in the cores of neutron stars. RHIC, which is slated to begin running in 1999 and to collect data for ten years, is the first systematic attempt to create this state of matter "in the lab".
The STAR (Solenoidal Tracker At RHIC) detector is designed to measure many quantities of interest per event, so that events can be categorized and correlated based on these characteristics. Nobody knows for sure how the presence of the QGP would manifest itself, so STAR was designed with a broad range of possible measurements in mind, to maximize the discovery potential. In addition, there are three other detectors at RHIC, built to carry out complementary as well as some overlapping measurements.
|
|
| Figure 1. A non-central event containing 1503 particle tracks. | Figure 2. A more complex event containing 8668 particle tracks. |
The figures show the outline of the STAR detector and the results of simulated gold-gold collision events at the maximum RHIC energy. This simulation was generated on the Pittsburgh Supercomputing Center (PSC) CRAY T3E massively parallel supercomputer, by means of the GSTAR software system developed by the STAR Collaboration using the CERN GEANT package. The figures illustrate the enormous complexity of each such event.
Considering that there will be up to 1000 collisions per second, of which STAR will only be able to save 1-2 for analysis, the magnitude of the online and offline computing challenge during the ten-year operation of STAR becomes evident. This is why up to two million events must be simulated before the experiment actually begins, allowing the STAR Collaboration to calibrate and tune their measurement and analysis systems. This process, called the Mock Data Challenge, has produced 250,000 events since July 1998, on the CRAY T3E machines at PSC and at the National Energy Research Supercomputing Center (NERSC) at Lawrence Berkeley National Laboratory (LBNL), as well as at the RHIC Computing Facility (RCF) at BNL.