Fraunhofer Institute for Industrial Mathematics ITWM

Fair

online / November 11, 2020 - November 19, 2020

Supercomputing 2020

International Conference for High Performance Computing, Networking, Storage, and Analysis

https://sc20.supercomputing.org/

SC has built a diverse community of participants including researchers, scientists, application developers, computing center staff and management, computing industry staff, agency program managers, journalists, and congressional staffers. This diversity is one of the conference's main strengths.

The technical program is the heart of SC. It has addressed various areas of scientific and engineering research, as well as technological development, innovation, and education with presentations, tutorials, panels, and discussion forums.

Our department "High Performance Computing" will present its latest research results at SC20.

What We Present at SC20

GaspiLS

GaspiLS is a scalable and industry proven linear solver library for the exascale age and trimmed to achieve optimal scalability.

GPI

GPI allows truly asynchronous and parallel communication of all threads and achieves optimal overlap of communication by computation.

GPI Space

GPI-Space allows the development of domain specific development- and runtime-systems.

Tarantella

Tarantella is an easy-to-use TensorFlow-based Deep Learning framework that enables scalable training of Deep Neural Networks on supercomputers.

BeeGFS

The parallel file system BeeGFS is used successfully in various complex environments around the world, in research as well as industry.

CARME

With the open source multi-user software stack Carme, several users can manage the available resources of a computing cluster.

EPI Processor

The European Processor Initiative (EPI) gets together 23 partners from 10 European countries, with the aim to bring to the market a low power microprocessor.

NASE

Neural Architecture Search Engine
AI-Services: Neural Architecture Search as a Service

ALOMA

ALOMA is a development platform for domain experts to integrate new and
existing algorithms and create complex workflows to be run on
large parallel systems.

DART

Based on our HPC distributed runtime environment called GPI-Space we provide an adjustable and highly efficient distributed runtime for Python applications. It inherites all features from GPI-Space in terms of fault-tolerance, flexibility, scalability and performance.