Issue No. 02 - March/April (2009 vol. 15)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TVCG.2008.85
Stephen Ingram , University of British Columbia, Vancouver
Tamara Munzner , University of British Columbia, Vancouver
Marc Olano , University of Maryland Baltimore County, Baltimore
We present Glimmer, a new multilevel algorithm for multidimensional scaling designed to exploit modern graphics processing unit (GPU) hardware. We also present GPU-SF, a parallel, force-based subsystem used by Glimmer. Glimmer organizes input into a hierarchy of levels and recursively applies GPU-SF to combine and refine the levels. The multilevel nature of the algorithm makes local minima less likely while the GPU parallelism improves speed of computation. We propose a robust termination condition for GPU-SF based on a filtered approximation of the normalized stress function. We demonstrate the benefits of Glimmer in terms of speed, normalized stress, and visual quality against several previous algorithms for a range of synthetic and real benchmark datasets. We also show that the performance of Glimmer on GPUs is substantially faster than a CPU implementation of the same algorithm.
Multivariate visualization, Information visualization
S. Ingram, M. Olano and T. Munzner, "Glimmer: Multilevel MDS on the GPU," in IEEE Transactions on Visualization & Computer Graphics, vol. 15, no. , pp. 249-261, 2008.