Hierarchical Clustering with CUDA/GPU.

Dar-Jen Chang, Mehmed Kantardzic, Ming Ouyang

2009 · 27 citations · 12 references

Concepts

Abstract

Graphics processing units (GPUs) are powerful com-putational devices tailored towards the needs of the 3-D gaming industry for high-performance, real-time graphics engines. Nvidia Corporation provides a programming lan-guage called CUDA for general-purpose GPU program-ming. Hierarchical clustering is a common method used to determine clusters of similar data points in multidimensional spaces; if there are n data points, it can be computed in O(n2) to O(n2 log n) sequential time, depending on the distance metrics employed. The present work explores paral-lel computation of hierarchical clustering with CUDA/GPU, and obtains an overall speed-up of up to 48 times over sequential computation with an Intel Pentium CPU. 1.

References

12