PARALLELIZING K-MEANS CLUSTERING USING GPU

Parallel Computing-Computer Engineering

The system with an optimized k-means implementation will be performed on the graphics processing unit (GPU). NVIDIA’s Computing Unified Device Architecture (CUDA), available from the G80 GPU family onwards, is used as the programming environment. Emphasis is placed on optimizations directly targeted at this architecture to best exploit the computational capabilities available. Clustering involves partitioning a set of objects into subsets called clusters so that objects in the same cluster are similar according to some metric. Clustering is widely used in many fields like machine learning, data mining, pattern recognition and bioinformatics. K-means clustering is very popular clustering method used which uses distance as the similarity measure. K-means algorithm uses a set of K random objects from the available data set and performs distance computations. K-means chooses multiple models with the goal of refinement and faster convergence. The focus of the proposed system is to investigate different approaches to parallelism and then ensemble K-means algorithm on modern many core hardware. The many core hardware involves GPUs and CPUs backed by the CUDA software stack. The main feature of the proposed system is that it makes use of the large computing capacity of the hardware by minimizing the amount of data access.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.