In nuclear and particle physics experiments, the beam intensities and rate of collisions are increasing at a dramatic rate; yet, the physics of interest is often contained in events that occur very rarely. This project will develop a trigger processor that can recognize these rare events in real-time, so as to greatly minimize required storage requirements and subsequent data processing. In particular, Field Programmable Gate Arrays (FPGAs) will be configured to efficiently identify the real-time vertex location in nuclear and high energy physics experiments. In Phase I, algorithms for line and vertex determination were developed and demonstrated for collisions expected at the Relativistic Heavy Ion Collider. A vertex finding algorithm was demonstrated on actual FPGA hardware, and architectural issues pertaining to the implementation of the trigger function were addressed. Phase II will build a prototype of an easily reconfigurable and scalable FPGA-based computing engine that will be suitable for the trigger processing function. The prototype will be used in demonstrations of computationally intensive commercial applications such as graphics processing, computer tomography, or real-time network traffic analysis.
Commercial Applications and Other Benefits as described by the awardee: The processor should be capable of trillions of operations per second with production costs as low as ten thousand dollars in a space comparable to a shoebox. In addition to applications in nuclear and high energy physics, the technology should contribute to commercial and governmental image processing problems such as synthetic aperture radar, computer tomography, beam-forming, image reconstruction and texturing, and code-breaking.
