Improving the performance for the range search on metric spaces using a multi-GPU platform

Abstract

Nowadays, similarity search is becoming a field of increasing interest because these kinds of methods can be applied to different areas in science and engineering, for instance, pattern recognition, information retrieval, etc. This search is carried out over metric indexes decreasing the number of distance evaluations during the search process, improving the efficiency of this process. However, for real applications, when processing large volumes of data, query response time can be quite high. In this case, it is necessary to apply mechanisms in order to significantly reduce the average query response time. In this sense, the parallelization of the metric structures processing is an interesting field of research. Modern GPU/Multi-GPU systems offer a very impressive cost/performance ratio. In this paper, we show a simple and fast implementation of similarity search method on a Multi-GPU platform. The main contributions are mainly the definition of a generic metric structure more suitable for GPU platforms, the efficient usage of GPU memory system and the implementation of the method in a Multi-GPU platform. © 2012 Springer-Verlag.