ScaDS Logo

CENTER FOR
SCALABLE DATA ANALYTICS
AND ARTIFICIAL INTELLIGENCE

News

👉 Stellenausschreibungen

Wir suchen drei studentische oder wissenschaftliche Hilfskräfte am Standort Leipzig. In folgenden Schwerpunkten liegen Ihre zukünftigen Tätigkeiten:

  • Einsatz von KI-Methoden am Anwendungsfall der additiven Fertigung

  • Document Layout Analysis und Recognition mit Machine Learning/KI

  • Frontend-Entwicklung für die entstehende Datenhandelsplattform DE4L

Was Sie erwartet?

  • Eine dynamisches, weltoffenes Team an einem von sechs deutschen KI-Forschungszentren
  • Sammeln Sie Erfahrung als Wissenschaftskommunikator*in
  • Unterstützen Sie den Ausbau des ScaDS.AI Dresden/Leipzig zu einem führenden deutschen Kompetenzzentrum für Big Data, Data Analytics und Künstliche Intelligenz

Wir freuen uns auf Ihr Know-how und Ihre Bewerbung!

Wir sehen uns in Leipzig

New HPC cluster taken into full operation

The new HPC cluster at the Center for Information Services and High Performance Computing (ZIH) at TU Dresden was taken into full operation. The cluster from NEC Deutschland GmbH will be used by our center for applications in the fields of Artificial Intelligence (AI). 

The TU Dresden is one of eleven German Universities of Excellence, a title which confirms the potential of one of Germany’s largest technical universities. TU Dresden offers a broad spectrum of study programmes, uniting the natural and engineering sciences with the humanities and social sciences, as well as medicine.

The new HPC cluster is especially designed for machine learning and was financed by the German Ministry for Science, Research and Education (BMBF) for the competence centre ScaDS.AI Dresden/Leipzig. After a Europe-wide tender, NEC Deutschland GmbH won the bid and installed the system already in December 2020.

At the heart of the system and essential for the computing power are a total of 272 NVIDIA A100 GPUs, eight of which are contained in each of the 34 compute nodes. Their theoretical maximum performance of floating point operations is more than 2.6 PFlop/s at 64-bit (double precision), more than 5.3 PFlop/s at 32-bit (single precision), and more than 42 PFlop/s in FT32-to-FP32 Tensor Operations. This is expected to make the system fast enough for an entry in the upcoming Top500 list in June 2021.

Each node also features a large 1 TB of main memory and 3.2 TB of local NVMe cache to quickly feed data to the GPUs. Fast connectivity to the central HPC storage complex is provided via two HDR InfiniBand ports each with a combined 400 Gbps of network bandwidth at a very low latency. The maximum power consumption of a node is 4.8 kW. Direct hot water cooling (DLC) ensures high energy efficiency while utilising the waste heat.

The new computing cluster will be integrated into the existing HPC infrastructure of the ZIH. As HPC competence centre, ZIH offers specialized computing resources as well as individual support and consulting for its users. The system will primarily be available for AI research of the competence centre ScaDS.AI Dresden/Leipzig. The execution of highly parallel applications that use AI methods for fast data analysis will benefit from this efficient system, driving both model development and expressiveness of analyses.

“The new Machine Learning solution from NEC provides us with a new level of High Performance Computing power for our AI research. The most important reasons for our decision have been the excellent computational capacity for the given budget, as well as a very convincing cooling concept. We are very impressed by the new system as well as by NEC’s excellence in providing service and support capabilities, and their expertise in AI” as Professor Dr. Wolfgang Nagel, Director at Center of Information Services and High Performance Computing explains.

“TU Dresden has an excellent reputation in research and ZIH is a very important HPC data centre in Germany. Therefore, we feel very honoured that NEC was given the task to deliver a new Petaflop system for their A.I. research,” Yuichi Kojima, Managing Director of NEC Deutschland GmbH and Vice President HPC at NEC Europe, adds.

----- 

About ZIH

ZIH is the university IT centre of TU Dresden and the High Performance Computing (HPC) competence centre for TU Dresden and the state of Saxony for over 20 years. Since the beginning of 2021, ZIH is one of eight NHR centres in the initiative “Nationales Hochleistungsrechnen” of the national Joint Science Conference (Gemeinsame Wissenschaftskonferenz, GWK) and the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) and thus offers its services to academic users from all over Germany.

 

About NEC Deutschland GmbH

NEC Deutschland GmbH is a wholly owned subsidiary of NEC Europe Ltd. and is a leading provider of HPC solutions, focusing on sustained performance for real-life scientific and engineering applications. To achieve this goal NEC delivers technology and professional services to industry and academia. Linux-based HPC clusters as well as our high-end vector systems meet the different needs of different customers in the most flexible way. Energy-efficiency is one of the key design objectives, addressed by advanced cooling technologies or by the high-bandwidth vector-architecture, which delivers unprecedented efficiency on real world code. The service capabilities from the operation of complex systems to the optimization of scientific codes and NEC's storage-appliances complete our solution offering.

Call for Papers: 2nd Workshop on Machine Learning on HPC Systems (MLHPCS) - in conjunction with ISC’21

The Workshop "Machine Learning on HPC Systems" (MLHPCS) is looking for your ideas!

MLHPCS ‘21 has a two stage submission process: submit a 1-2 page extended abstract to contribute with a talk to MLHPCS. Additionally, authors optionally can submit a post-conference paper for publication in LNCS!

Over the last few years, Machine Learning (and in particular Deep Learning) (ML / DL) has become an important research topic in the High Performance Computing (HPC) community. This comes along with new users and data intensive applications on HPC systems, which increasingly affects the design and operation of compute infrastructures. Bringing new users and data intensive applications on HPC systems, Learning methods are increasingly affecting the design and operation of compute infrastructures. On the other hand, the learning community is just getting started to utilize the performance of HPC, leaving many opportunities for better parallelization and scalability. 

Important Dates:

→  Extended Abstracts deadline: May 14th

 

The intent of this workshop is to bring together researchers and practitioners from all communities to discuss three key topics in the context of High Performance Computing and learning methods:

  • parallelization and scaling of ML / DL algorithms,
  • learnig applications on HPC systems, and
  • HPC systems design and optimization for ML / DL workloads.

 

Topics / Scope:

  • Unsolved problems in ML / DL on HPC systems
  • Scalable Machine Learning / Deep Learning algorithms
  • Parallelization techniques
  • Libraries for ML / DL
  • Tools + workflows for ML / DL on HPC systems
  • Optimized HPC system design / setup for efficient ML / DL
  • ML Applications on HPC Systems

 

MLHPCS will be held as a live online workshop within the ISC conference on 02/07/2021.

Post-Conference papers will be published in Springer LNCS.

More information

 

New Machine Learning Cluster for ScaDS.AI

In February, a new HPC cluster, which is especially designed for machine learning, will be handed over to user operation at the data center of TU Dresden (LZR - Rechenzentrum Lehmann-Zentrum). It was financed from a special budget of the BMBF for our competence center ScaDS.AI Dresden/Leipzig.

After a Europe-wide tender, NEC Deutschland GmbH won the bid and was able to start installation in December 2020. At the heart and essential for the computing power are a total of 272 A100 GPUs from NVIDIA. Eight of these GPUs are contained in each of the 34 compute nodes. Their respective theoretical maximum performance of floating point operations is more than 2.6 PFlop/s at 64-bit and more than 5.3 PFlop/s at 32-bit. This is expected to make the system fast enough for an entry in the upcoming Top500 list in June 2021.

Each node also features a large 1 TB of main memory and 3.2 TB of local NVMe cache to quickly feed data to the GPUs. Fast connectivity to the central HPC storage complex is provided via two HDR Infiniband ports each with a combined 400 Gbps. The maximum power consumption of a node is 4.8 kW. Direct hot water cooling ensures high energy efficiency while utilizing the waste heat.

The new computing cluster will be integrated into the existing high-performance computing infrastructure of the Center for Information Services and High Performance Computing (ZIH) and will primarily be available for the research of the competence center ScaDS.AI Dresden/Leipzig. In particular, the execution of highly parallel applications that use artificial intelligence methods for fast data analysis will benefit from this efficient system and advance both model development and expressiveness of analyses. Currently, the system is in the acceptance phase and is available for initial testing with scientific applications. Access can be requested via an HPC-DA project application on the ZIH HPC web pages: https://tu-dresden.de/zih/hochleistungsrechnen/zugang

🤖 Podcast-Beitrag: Neustart – Die Zukunft beginnt mit uns 🤖

Episode 5: Die Macht der der Algorithmen

Unser Leipziger Forscher Dr. Christian Martin diskutiert in der aktuellen Folge von Neustart-Die Zukunft beginnt mit uns die Chancen, Möglichkeiten, aber auch Risiken algorithmischer Verfahren gemeinsam mit dem Moderator Tobias Hülswitt und der Politikwissenschaftlerin Prof. Dr. iur. Sabine Müller-Mall von der TU Dresden. Was kann ein Algorithmus wirklich und was kann er nicht? Welche Gefahren entstehen durch algorithmischen Voreingenommenheit (Bias) und inwieweit ist eine solche Voreingenommenheit diskriminierend?

Die Episode kann unter Podcast oder Apple Podcast gehört und die Show abonniert werden. 

 

Podcast: https://www.podcast.de/episode/525093177/DIE+MACHT+DER+ALGORITHMEN/

Apple Podcast: https://podcasts.apple.com/ca/podcast/die-macht-der-algorithmen/id1541161197?i=1000506045271