A Comparative Analysis of Algorithmic Music Generation on GPUs and FPGAs

The authors in this paper aim towards comparing the efficiency of Music Generation with Recurrent Neural Networks on GPUs and FPGAs. A more common approach towards Deep Learning has always been the GPUs. However, FPGAs are a new addition to the field of deep learning computations. This paper is aimed at the developing a deep learning module based on Recurrent Neural Networks for algorithmic music generation. It therefore, compares the efficiency of this algorithm on FPGAs and GPUs for finding the best suitable hardware for obtaining maximum efficiency (Abstract)