A GPU-Accelerated Large-Scale Music Similarity Retrieval Method

High-quality content-based music similarity retrieval methods are non-vectorial and use non-metric divergence measures, which prevents the expansion of music recommendation systems. We presents a GPU-based method to speed up content-based music similarity search in large-scale collections, in order to improve the response speed without reducing retrieval accuracy. The method also introduce an optimization technique based on memory layout to improve memory access. The efficiency of our method is validated through extensive experiments. Evaluation results show that our single GPU implementation achieves 10x speedup ratio on NVIDIA GTX480, when compared to a typical general purpose CPU´s execution time.