News

2024

16.05.2024 - HPC|A Group and NHR@SW at ISC High Performance 2024

A successful week at ISC High Performance 2024 is coming to an end! As a member of the NHR@SW center, the Johannes Gutenberg University Mainz was represented, among other things, at the NHR Alliance's booth. The research of the HPC|A group could be found at various exhibition booths. Fraunhofer ITWM, for example, showcased Niklas Bartelheimer's work about the GPI-2 BXI integration. During the workshop day, Sarah Neuwirth presented the "Tracking the Data Trail" project at the HPC I/O in the Data Center (HPC-IODC 2024) Workshop, which is a joint effort with Hariharan Devarajan from LLNL and Jay Lofstead from SNL.

08.05.2024 - REX-IO Workshop accepted at IEEE Cluster 2024

Great news! The 4th edition of the REX-IO Workshop (Re-envisioning Extreme-Scale I/O for Emerging Hybrid HPC Workloads) has been accepted as a full-day workshop at IEEE Cluster 2024! REX-IO 2024 is jointly organized with Arnab K. Paul from BITS Pilani, K K Birla Goa Campus, India and Jay Lofstead from Sandia National Laboratories, USA and will be held on September 24, 2024 in conjunction with IEEE Cluster 2024 in Kobe, Japan. Stay tuned!

25.04.2024 - Invited Guest Lecture in TOPML Lecture Series

The interdisciplinary research project TOPML (Trading off Non-Functional Properties of Machine Learning) investigates the complex interactions and dependencies between various non-functional requirements for machine learning, such as transparency and data security. The lecture series "Transparenz, Datensicherheit & Co.: Umsetzbare Anforderungen an KI-Systeme?" in the summer semester 2024, organised by TOPML, is aimed at the academic community and interested students. Each topic will be examined by two speakers from both a research and a practical perspective and will provide space for further discussion afterwards. Together with Dr. Enver Solan from GEMESYS GmbH, Prof. Neuwirth will hold the guest lecture on "Ressourcen-Effizienz" on Tuesday, May 21, 2024. Further information can be found on the TOPML event website.

16.04.2024 - New Group Member

We are happy to welcome our new group member Martin Machajewski! He recently graduated with a M.Sc. in Computer Science from Goethe University Frankfurt where he focused on the development of an efficient monitoring infrastructure of node-level statistics and resource utilization in HPC systems. Martin will continue his work on holistic and modular performance monitoring and modeling for HPC systems as part of his PhD. He will also support the NHR South-West center in the development of Performance Engineering Method Labs.

08.03.2024 - Keynote at 1st Proud and Strong in Computing Conference (PSCC)

Sarah Neuwirth has been invited to give the opening keynote talk at the first Proud and Strong in Computing Conference (PSCC) to be held from June 24 to 25, 2024 in Goettingen, Germany. The event will focus on raising awareness for the (lack of) diversity in scientific computing and related disciplines as well as networking. There, Sarah will share both her personal and professional journey in her keynote titled "Empowering Change: A Personal Story of Breaking Barriers in HPC".

06.03.2024 - Invited Talk at ACM PASC'24 Conference 

Sarah Neuwirth has been invited to give a talk at the ACM Platform for Advanced Scientific Computing (PASC'24) Conference to be held from June 3 to 5, 2024 in Zurich, Switzerland. The minisymposium "Breaking the HPC Silos for Outstanding Results" will host three scientific presentations from HPC experts from different backgrounds, which offer a view on a variety of carrier paths, focusing on how inclusivity and diversity played a role in achieving the presented results. There, Sarah will share her experience in pushing performance of applications to reach new horizons in her talk titled "Breaking Limits: Scaling HPC Performance Engineering Horizons to Maximize Potential".

01.03.2024 - New Group Member

We are happy to welcome our new group member Niklas Bartelheimer! He holds a M.Sc. in Computer Engineering from Heidelberg University and has extensive knowledge in low-level network interfaces, including Ethernet, Infiniband and RDMA over Converged Ethernet (RoCE). Niklas will continue his work in the EUPEX project and his Ph.D. in Computer Science at JGU Mainz with a focus on parallel programming models, the BXI interconnect and direct communication methods for distributed accelerator technologies. 

28.02.2024 - Call for Papers for the WHPC Session at Euro-Par 2024

Jointly organized with Marta Garcia-Gasulla from Barcelona Supercomputing Centre, the HPC|A group is happy to invite paper submissions to the 1st WHPC Session to be held in conjunction with the 30th International European Conference on Parallel and Distributed Computing (Euro-Par 2024) from 26 to 30 August, 2024 in Madrid, Spain. Euro-Par is the prime European conference covering all aspects of parallel and distributed processing, ranging from theory to practice, from small to the largest parallel and distributed systems and infrastructures, from fundamental computational problems to full-fledged applications, from architecture, compiler, language and interface design and implementation, to tools, support infrastructures, and application performance aspects. 

New to the Euro-Par technical program, the WHPC Session is dedicated to women who are students or in the early stages of their HPC careers, typically within the first five years after graduation or via a career change to the field of HPC. Papers accepted for the WHPC session will be published in the Euro-Par 2024 companion proceedings. The WHPC session will include technical presentations to promote and showcase their research and is complemented by a keynote speech. Papers can be submitted until May 6, 2024 11:59pm AoE. 

14.02.2024 - EOFS Workshop on Open Parallel File Systems co-organized by HPC|A Group

Co-organized by EOFS, JGU Mainz, TU Dresden and University of Hamburg, we are proud to announce the final agenda of the Workshop on "Offene Parallele Filesysteme: Status, Entwicklungstrends und Aspekte der Informatikausbildung". The second edition of this workship will take place on February 22-23, 2024 at TU Dresden. The aim of the workshop is to exchange positive and negative experiences and concepts in research and teaching and to assess development trends in the German-speaking community.

01.02.2024 - New Group Member

We are happy to welcome our new group member Zhaobin Zhu! In 2021, he graduated with a M.Sc. in Applied Computer Science from Heidelberg University, where he focused on the simulation and analysis of high-performance on-chip networks, before joining Goethe University Frankfurt as a Ph.D. candidate. Zhaobin will continue his Ph.D. in Computer Science at JGU Mainz with a focus on parallel I/O optimization, reproducible benchmarking, and modular and automated workload analysis for HPC systems.

12.01.2024 - SRA6 "I/O & Storage" Working Group Presentation at ETP4HPC Conference 2024

The second edition of the ETP4HPC Conference will take place on 13-14 February 2024 at the Van der Valk Hotel in Sassenheim-Leiden (near Amsterdam). On the second conference day, SRA6 will present its main research directions and interaction with the RIAG/EuroHPC JU/EC, and will hold an SRA6 workshop where the hardware and software working groups will present and discuss their progress.

04.01.2024 - Invited Talk at SOS26 Workshop

SOS is an invitation-only series of highly interactive workshops on distributed supercomputing organized by the Association for High Speed Computing with a program developed by Sandia National Laboratories, Oak Ridge National Laboratory, and the Swiss National Supercomputing Center. The theme of the 26th edition of the SOS workshop (SOS26) will be “Versatile HPC for the evolving and expanding needs of science”. SOS26 will be held in Cocoa Beach, FL, USA from March 11 to 14, 2024. Sarah Neuwirth has been invited as a speaker and panelist for the "Reproducibility and Data Management" session. In this session, she will present her work in a talk titled "Bridging the Gap: Parallel Storage, Monitoring Tools, and the Quest for Reproducibility?".

02.01.2024 - ESSA Workshop at IEEE IPDPS 2024

Jointly with colleagues from Inria (France), Florida State University (USA) and BITS Pilani K K Birla Goa Campus (India), Sarah Neuwirth organizes the 5th Workshop on Extreme-Scale Storage and Analysis (ESSA 2024), to be held on May 27, 2024 in conjunction with IEEE IPDPS 2024 in San Francisco, CA, USA. The ESSA Workshop aims to bring together experts, researchers and developers in data-related areas including storage, I/O, processing and analysis on extreme scale infrastructures including HPC systems, clouds, edge systems or hybrid combinations of those, to discuss advances and possible solutions. The call for papers will accept submissions until February 15, 2024 AoE!

2023

26.12.2023 - Research Paper accepted at ACM Transactions on Storage

An early New Year's gift! Our research paper titled "Tarazu: An Adaptive End-to-End I/O Load Balancing Framework for Large-Scale Parallel File Systems" has been accepted at one of the best journals in systems in Computer Science - ACM Transactions on Storage (TOS)! The paper is a testament that hard team work and quality research surely find their way to publication. After multiple downs (3 major revisions, 1 journal change, and 1 minor revision), our paper finally sees the light of day. Despite our busy schedules and the different time zones, we devoted many hours to this work. Arnab K. Paul from BITS Pilani KK Birla Goa Campus and I are truly proud that our work will be featured in ACM TOS.

30.11.2023 - Invited Talk in the ZIH-Colloquium Series at TU Dresden

The ZIH colloquium is a public event that usually takes place on the 4th Thursday of each month. It is organized by the Center for Information Services and High Performance Computing (ZIH) at TU Dresden. On November 30, 2023, Sarah Neuwirth has been invited to talk about "Holistic HPC Performance Engineering and Reproducible Benchmarking: Opportunities and Challenges".

24.11.2023 - Paper accepted in Special Issue on APDCM 2023 to appear in January 2024 Issue of IJNC

Congratulations to Niklas Bartelheimer for getting his paper accepted in the Special Issue on APDCM 2023 in the International Journal of Networking and Computing (IJNC). After presenting initial results in his short paper titled "Toward a Modular Workflow for Network Performance Characterization" at the 25th Workshop on Advances in Parallel and Distributed Computational Models (APDCM'23), Niklas submitted a full version to the Special Issue on APDCM 2023. His paper titled "Automated Network Performance Characterization for HPC Systems" was subject to a rigorous peer review process and will be published in the January 2024 issue of IJNC.
Update (07.01.2024): Read the full paper now online on IJNC!

23.11.2023 - ETP4HPC SRA6 Working Group "I/O & Storage" led by HPC|A Group

ETP4HPC – the European Technology Platform (ETP) for High-Performance Computing (HPC) - is a private, industry-led and non-profit association. Its main mission is to promote European HPC research and innovation in order to maximize the economic and societal benefit of HPC for European science, industry and citizens. ETP4HPC's main task is to propose research priorities and programme contents in the area of HPC technology and usage, by issuing the Strategic Research Agenda (SRA) periodically. The SRA is used by the EuroHPC Joint Undertaking (JU) to define the contents of the HPC Technology Work Programmes. For SRA6 (release planned for early September 2024), the HPC|A Group will lead the Working Group "I/O & Storage".

06.11.2023 - Invited Talk at Sandia National Laboratories

The "Scalable System Software" Group at Sandia National Laboratories has invited Prof. Neuwirth to present her latest research at the New Mexico campus in Albuquerque. Sarah will talk about "Characterizing HPC I/O Workloads through Holistic and Reproducible Performance Engineering".

02.11.2023 - NHR-Computational Physics Symposium 2023

The HPC|A group is happy to announce the two-day NHR-Computational Physics Symposium 2023, which will take place on November 2/3 online. The symposium is co-organized by PC2, NHR@KIT and NHR@SW (represented by JGU), and is an initiative of the computational physics center of the National Alliance of High-Performance Computing (NHR), which bundles the resources and competences of high-performance computing and makes them available to scientists of German universities. 

Experts from different fields, e.g. condensed matter physics, high energy physics and astroparticle physics, will present their research goals and numerical challenges in high performance computing. Further details, the (preliminary) timetable, and the registration can be found at: https://pc2.de/go/nhr-computational-physics-2023

31.10.2023 - REX-IO Workshop at IEEE Cluster 2023

Jointly with colleagues from BITS Pilani, K K Birla Goa Campus, India and Sandia National Laboratories, USA, the High Performance Computing and its Applications (HPC|A) Group organizes the 3rd edition of the REX-IO Workshop (Re-envisioning Extreme-Scale I/O for Emerging Hybrid HPC Workloads), to be held on October 31, 2023 in conjunction with IEEE Cluster 2023 in Santa Fe, NM, USA. REX-IO solicits novel work that characterizes I/O behavior and identifies the challenges in scientific data and storage management for emerging HPC workloads, introduces potential solutions to alleviate some of these challenges, and demonstrates the effectiveness of the proposed solutions to improve I/O performance for the exascale supercomputing era and beyond.

27.10.2023 - Two Research Papers accepted at IEEE ICPADS 2023

Congratulations to Zhaobin Zhu and Niklas Bartelheimer for the acceptance of their research papers at the 29th IEEE International Conference on Parallel and Distributed Systems (ICPADS 2023) to be held from December 17 to 21, 2023, in Ocean Flower Island, Hainan, China. With an unexpected record number of 1008 submissions, the competition was truly fierce this year, and all papers were subject to a rigorous peer review process, with less than 19% of submissions accepted for publication. Zhaobin will present his work titled "Characterization of Large-Scale HPC Workloads With Non-Naïve I/O Roofline Modeling and Scoring" and Niklas will present his paper titled "Toward Reproducible Benchmarking of PGAS and MPI Communication Schemes". Once again, congratulations!

25.10.2023 - Invited Talk organized by Gesellschaft für Informatik (GI) Regionalgruppe Rhein-Main

Prof. Neuwirth has been invited to present her research work at the next event organized by GI Regionalgruppe Rhein-Main on October 25, 2023. The topic of her talk is "HPC Performance Engineering im Zeitalter von Exascale-Systemen und Künstlicher Intelligenz". The event will take place at the Hörsaal des Instituts für Kernphysik, Johann-Joachim-Becher-Weg 45. Further information and the complimentary registration can be found here.

10.10.2023 - Invited Talk at the 5th Annual Meeting of M3ODEL

The 5th Annual Meeting of M3ODEL (Mainz Institute of Multiscale Modeling) took place from October 9 to October 11 at the Helmholtz Institute Mainz. Here, Prof. Neuwirth presented her work in an invited talk titled "Unleashing the Potential of Exascale Scientific Research through HPC Performance Engineering". 

01.10.2023 - New research group established

HPC researcher and system engineer Prof. Dr. Sarah Neuwirth is coming to Johannes Gutenberg University Mainz. Starting October 1, 2023, she will fill the newly created professorship for "High Performance Computing and its Applications" at the Institute for Computer Science. Sarah's responsibilities will also include the management and further development of JGU's High Performance Computing division, the coordination of HPC activities at regional and national levels and the representation of JGU in the NHR, the NHR South-West, and the Gauss-Allianz.