The evolution of High-Performance Computing (HPC) platforms enables the design and execution of progressively larger and more complex workflow applications in these systems. The complexity comes not only from the number of elements that compose the workflows but also from the type of computations they perform. While traditional HPC workflows target simulations and modelling of physical phenomena, current needs require additional data analytics (DA) and artificial intelligence (AI) tasks. However, these workflows are not just complex to develop, they are also difficult to deploy and execute in shared and restricted HPC sites. Workflow developers implement tools to orchestrate the deployment, configuration and execution of different frameworks which are used in the different workflow steps (HPC simulations or DA and AI algorithms).
In this talk, the speaker will present the problems and challenges to facilitate the development, deployment, and execution of these complex workflows in the HPC sites and their analogies with the Function-as-a-Service model used in the Cloud Computing environment. He will introduce the methodology developed by the EuroHPC eFlows4HPC project which focuses on how similar approaches can be followed to simplify the deployment and execution of Complex workflows in HPC systems.
Jorge Ejarque holds a PhD on Computer Science (2015) a MSc. on Computer Architecture Networks and Systems (2009) and an engineering degree on Telecommunications (2005) from the Technical University of Catalonia (UPC). In 2005, he worked as IT consultant in Better Consulting. Since the end of 2005 to 2008, he worked as research support engineer at UPC and at the end of 2008, he joined to the Grid Computing group at BSC.
During his career at the BSC, he has contributed in the design and development of different tools and programming models for HPC in distributed platforms. He has been involved in several National and European R&D projects. He has been member of a program committee of several international conferences and reviewer of journal articles. He was a member of the Spanish National Grid Initiative panel. His current research interests are focused on introducing energy efficiency in parallel programming models for heterogeneous parallel distributed computing environments and semantic interoperability between distributed computing platforms.