Parallel I/O analysis in distributed deep learning applications on high-performance computing | Publicación