Speaker: Rosa M. Badia, Barcelona Supercomputing Center

Moderator: Esteban Mocskos, Universidad de Buenos Aires

The evolution of High-Performance Computing (HPC) systems towards every-time more complex machines is opening the opportunity of hosting larger and heterogeneous applications. In this sense, the demand for developing applications that are not purely HPC, but that combine aspects of Artifical Intelligence and or Data analytics is becoming more common. However, there is a lack of environments that support the development of these complex workflows. The webinar will present PyCOMPSs, a parallel task-based programming in Python. Based on simple annotations, sequential Python programs can be executed in parallel in HPC-clusters and other distributed infrastructures.

PyCOMPSs has been extended to support tasks that invoke HPC applications and can be combined with Artificial Intelligence and Data analytics frameworks.

Some of these extensions are made in the framework of the eFlows4HPC project, which in addition is developing the HPC Workflows as a Service (HPCWaaS) methodology to make the development, deployment, execution and reuse of workflows easier. The webinar will present the current status of the PyCOMPSs programming model and how it is being extended in the eFlows4HPC project towards the project needs. Also, the HPCWaaS methodology will be introduced. 

About the speaker: Rosa M. Badia holds a PhD on Computer Science (1994) from the Technical University of Catalonia (UPC).  She is the manager of the Workflows and Distributed Computing research group at the Barcelona Supercomputing Center (BSC).

Her current research interests are programming models for complex platforms (from edge, fog, to Clouds and large HPC systems).  The group led by Dr. Badia has been developing StarSs programming model for more than 15 years, with a high success in adoption by application developers. Currently the group focuses its efforts in PyCOMPSs/COMPSs, an instance of the programming model for distributed computing including Cloud.

Dr Badia has published nearly 200 papers in international conferences and journals in the topics of her research. Her group is very active in projects funded by the European Commission and in contracts with industry. Dr Badia is the PI of the eFlows4HPC project.

Registrations are now closed.

Speakers: Vitor Silva and Filipe Guimarães, Jülich Supercomputer Centre

Moderator: Esteban Mocskos, Universidad de Buenos Aires

Check the speakers’ presentation slides here. 

LLview is a monitoring infrastructure developed by the Jülich Supercomputing Centre with the objective to provide an easy to use and adaptable software suite for monitoring High Performance Computing systems. With the emergence of large heterogeneous machines, in the range of Exascale, the challenges of monitoring such huge systems increase significantly. To address that, LLview is under continuous development in order to work for a wide range of hardware systems and software interfaces with negligible overhead and at the same time providing fast, reliable access to job reports, system-wide monitoring data, and real-time system information. That information is provided to system users, project advisors, support teams and system administrators, helping the managing of jobs, identification of performance issues at many levels and also helping the system administrators to find failures and system malfunctions. This webinar gives an overview of the different LLview components and their interaction with each other and the system. Moreover, particular attention is drawn to the system monitoring views and the job reporting features, as they allow to trace the entire life cycle of a job and can help identify problems and bottlenecks at a very early stage.

About the Speakers:

Vitor Silva received his Computer Science degree from Universiade Federal de Minas Gerais. His M.Sc was earned in Systems and Computer Engineering from Universidade Federal do Rio de Janeiro and later received his Ph.D from Universidade Federal de Minas Gerais, this time in Nuclear Engineering. He worked as software developer in the digital image processing field, but most of his career was in the Nuclear Engineering field, mainly working with computer modeling and solving Neutronics and Thermal-hydraulics problems related to nuclear reactors. He was also the main admin of a small cluster system installed from scratch. Since 2021 he has been working at the Jülich Supercomputing Centre with monitoring tools and simulation.

Filipe Guimarães is a computational physicist. Graduated in Physics, M.Sc in Physics and Ph.D in Physics from the Universidade Federal Fluminense. He has been working with High Performance Computing since 2014 – initially from a user’s side, but moved to the support side in 2020. Since then, one of his focuses was to improve monitoring tools used and developed at the Jülich Supercomputing Centre.

About the Moderator: Esteban Mocskos is a full-time professor at Universidad de Buenos Aires (UBA) and researcher at the Center for Computer Simulation (CSC-CONICET). He received his Ph.D. in Computer Science from UBA in 2008 and was postdoc at the Protein Modelling group at UBA. His research interests include distributed systems & blockchain, computer networks, processor architecture, and parallel programming. He is part of the steering committee of the Latin-American HPC CARLA conference and onE of the committee members of Argentina’s National HPC system.

Speaker: Marc-André Hermanns, RWTH Aachen

Moderator: Bernd Mohr, Jülich Supercomputer Centre

JUBE can help in the automating application benchmarking on a given platform. JUBE’s features in automatic sandboxing and parameter-space creation can assist to easily sweep build and runtime parameters for an application on a given platform to identify the best build and run configuration.

This talk provides some lessons learned in building a JUBE-based benchmark Suite for the RWTH Aachen University Job-Mix that reduces redundancy of information and allows for easy integration of future applications. It will specifically address advanced features for parameter settings, parameter inheritance, and some tips and tricks to overcome some of its limitations.

About the speaker: Marc-André Hermanns is a member of the HPC group at the IT Center of RWTH Aachen University. His research focuses on tools and interfaces for the performance analysis of parallel applications. He has been involved in the design and implementation of various courses on topics of parallel programming for high-performance computing. Next to supporting HPC users as part of the competence network for high-performance computing in North-Rhinewestphalia (HPC.NRW), he also contributes to the development of online tutorials and courses within the competence network. He is a long time user and advocator for JUBE and created configurations for various applications and benchmarks, both for classical system benchmarking, as well as integration of performance analysis tools in such workflows.

About the moderator: Bernd Mohr started to design and develop tools for performance analysis of parallel programs already with his diploma thesis (1987) at the University of Erlangen in Germany, and continued this in his Ph.D. work (1987 to 1992).  During a three year postdoc position at the University of Oregon, he designed and implemented the original TAU performance analysis framework. Since 1996 he has been a senior scientist at Forschungszentrum Juelich. Since 2000, he has been the team leader of the group ”Programming Environments and Performance Analysis”. Besides being responsible for user support and training in regard to performance tools at the Juelich Supercomputing Centre (JSC), he is leading the Scalasca performance tools efforts in collaboration with Prof.  Felix Wolf of TU Darmstadt. Since 2007, he has also served as deputy head for the JSC division ”Application support”. He was an active member in the International Exascale Software Project (IESP/BDEC) and work package leader in the European (EESI2) and Juelich (EIC, ECL) Exascale efforts.  For the SC and ISC Conference series, he served on the Steering Committee.  He is the author of several dozen conference and journal articles about performance analysis and tuning of parallel programs.

Registrations are now closed.

Speaker: Jens Henrik Göbbert, JSC

Moderator: Esteban Mocskos, Universidad de Buenos Aires

Interactive exploration and analysis of large amounts of data from scientific simulations, in-situ visualization and application control are convincing scenarios for explorative sciences. Based on the open source software Jupyter or JupyterLab, a way has been available for some time now that combines interactive with reproducible computing while at the same time meeting the challenges of support for the wide range of different software workflows.

Even on supercomputers, the method enables the creation of documents that combine live code with narrative text, mathematical equations, visualizations, interactive controls, and other extensive output. However, a number of challenges must be mastered in order to make existing workflows ready for interactive high-performance computing. With so many possibilities, it’s easy to lose sight of the big picture. This webinar provides a compact introduction to high performance interactive computing.

About the Speaker: Jens Henrik Göbbert graduated in mechanical engineering in 2006 and worked until 2014 as a research assistant at the Institute for Technical Combustion in the area of turbulence modelling and high performance computing. He joined the cross-sectional group “Immersive Visualization” of the Jülich Aachen Research Alliance (part of the Virtual Reality Group of the IT Center at the RWTH Aachen University) and became part of the cross-sectional team “Visualization” of the Jülich Supercomputing Center at the FZJ in 2016 as an expert in visualization of large scientific data sets, in situ visualization & coupling and interactive supercomputing.

About the Moderator: Esteban Mocskos is a full-time professor at Universidad de Buenos Aires (UBA) and researcher at the Center for Computer Simulation (CSC-CONICET). He received his Ph.D. in Computer Science from UBA in 2008 and was postdoc at the Protein Modelling group at UBA. His research interests include distributed systems & blockchain, computer networks, processor architecture, and parallel programming. He is part of the steering committee of the Latin-American HPC CARLA conference and one of the committee members of Argentina’s National HPC system.