Approach

theHIVE is a Beowulf-class parallel computer consisting of 66 Pentium Pro PC's (64 slaves and 2 controllers) with 2 processors per PC (for 128 total processors). It was developed and assembled by the Applied Information Sciences Branch at GSFC.

Accomplishments

A variety of applications are running on theHIVE. Two significant results from this past year are as follows:

1. A computationally intensive, recursive image segmentation approach was implemented on theHIVE and the 512-processor ESS CRAY T3E using the C programming language and PVM (Parallel Virtual Machine) software. The image segmentation algorithm, which is designed for the analysis of multispectral or hyperspectral remotely sensed imagery data, is a hybrid of region growing and spectral clustering that produces a hierarchy of image segmentation based on detected natural convergence points. Timing results on Landsat Multispectral Scanner data show, that for moderately large images (about 2000x2000 pixels), the algorithm runs about 1.3 times faster on the HIVE, even though the HIVE is some 86 times less costly than the Cray T3E.

2. Researchers surveyed 96 exotic cosmological models containing both massive neutrinos and a cosmological constant, using theHIVE to implement an embarrassingly parallel parameter search. These models are then compared to the measured fluctuation spectrum at low redshift, and microwave background fluctuations on scales of around one degree, to pick the most promising set of parameters. They find two moderately promising models, one with relatively high density (60 percent of critical) and one with relatively low density (40 percent), but nothing compelling. The researchers believe the results from this study are the first scientific results to come from Goddard's Beowulf computers.

Points of Contact

John Dorband
NASA Goddard Space Flight Center
John.E.Dorband.1@gsfc.nasa.gov
301 286-9419

James Tilton
NASA Goddard Space Flight Center
James.C.Tilton.1@gsfc.nasa.gov
301-286-9510

http://newton.gsfc.nasa.gov/thehive/