Computational & Technology Resources
an online resource for computational,
engineering & technology publications
PROCEEDINGS OF THE SIXTH INTERNATIONAL CONFERENCE ON PARALLEL, DISTRIBUTED, GPU AND CLOUD COMPUTING FOR ENGINEERING
Edited by: P. Iványi and B.H.V. Topping
High Performance Computing and Machine Learning
High Performance Computing Center Stuttgart (HLRS), University of Stuttgart, Germany
M.M. Resch, "High Performance Computing and Machine Learning", in P. Iványi, B.H.V. Topping, (Editors), "Proceedings of the Sixth International Conference on Parallel, Distributed, GPU and Cloud Computing for Engineering", Civil-Comp Press, Stirlingshire, UK, Paper 1, 2019. doi:10.4203/ccp.112.1
Keywords: high performance computing, machine learning, artificial intelligence.
Machine Learning (ML) and Artificial Intelligence (AI) are fields that show a huge potential for using computers in a variety of fields. Areas of application range from improving and speeding up medical image processing to optimizing urban planning processes to a standardized and high speed handling of banking processes even in the usually heavily personalized consumer market. Some economists assume that ML and AI will change the world so dramatically that millions of jobs will be lost and we need to speak of a “second machine age”.
In this article we have a look at the merger of High Performance Computing (HPC) with ML and AI. The situation of HPC has been described before and is considered to be interesting but also limited by the technical problems that we face with the end of Moore’s law . We will argue that ML and AI have to be seen as two different technologies that are part of a chain of technologies that naturally lead us from HPC to AI. We will also argue that HPC is both a facilitating technology for ML and AI and a heavy user of ML and AI technologies. In this paper we will not address ethical issues that come with ML and especially AI. This is an important topic but should not be part of a technical overview paper.
Finally, we give an outlook of how HPC, ML and AI will be merged into a new world of technologies. This merger will allow to tackle new research questions and will help to substantially improve our way of doing simulation on supercomputers.
purchase the full-text of this paper (price £22)