A digital signal decomposition algorithm for GRETINA

Energy-tracking arrays such as GRETINA and AGATA will be the next generation of detector systems for in-beam gamma-ray spectroscopy. They depend on digital pulse processing to extract the positions and energies of multiple interactions in segmented high-purity Ge detectors. This processing makes use of digitized signals both from the hit segments themselves, and the induced image charges in neighboring segments. The complex algorithms involved make this process the major computational bottleneck for these detectors, requiring a computer farm at least of the order of 100 high-speed processors. For GRETINA, candidate algorithms for signal decomposition include Adaptive Grid Search, Singular Value Decomposition, and Sequential Quadratic Programming. Current work on the development of the algorithm, and verification of its efficacy and resolution, will be described.