building for unpacking and pre-assembling racks for new supercomputers

MPCDF picture of the month: January 2025

We are currently building an extension to our technical building. Spanning 100 square metres, it will primarily be used for unpacking and pre-assembling racks for new supercomputers. This will help to keep flammable packing materials outside of the computer halls, thereby enhancing fire safety. It will also prevent dirt and dust from reaching the operational systems. The new extension will provide additional storage space for spare parts and other materials. Its completion is expected early this year.

High-performance computing and data analytics application support for the MPG

High-performance computing and data analytics application support for the MPG

The MPCDF provides high-level support for the development, optimization, analysis and visualization of high-performance computing (HPC) and data analytics (HPDA) applications to Max-Planck Institutes with high-end computing needs, e.g. in astrophysics, fusion research, materials and bio sciences, polymer research, and theoretical chemistry.

Viper-CPU (deployment at LRZ). Image credit: H. Huber (LRZ)

MPG supercomputer Viper

Viper-CPU: Based on AMD EPYC Genoa 9554 processors: 768 CPU compute nodes, 98304 CPU cores, 432 TB RAM (DDR5), 4.9 PFlop/s theoretical peak performance (FP64). 200 Gb/s Interconnect (Nvidia NDR200, nonblocking fat-tree fabric over all nodes). 4.3 PFlop/s HPL performance. Operational since June 2024.
 

Viper-GPU (deployment at MPCDF)

MPG Supercomputer Viper

Viper-GPU: Based on AMD Instinct MI300A APUs: 300 nodes, 600 MI300A APUs, 75 TB RAM (HBM3). 400 Gb/s Interconnect (Nvidia NDR, nonblocking fat-tree fabric over all nodes). 31.1 PFlop/s HPL performance. Operational since February 2025.

MPG supercomputer Raven

MPG supercomputer Raven

Based on Intel Xeon IceLake-SP processor 8360Y and Nvidia A100 GPU: 1592 CPU-compute nodes, 114,624 CPU-cores, 457 TB RAM (DDR4), 8.8 PFlop/s theoretical peak performance (FP64), 192 GPU-accelerated nodes providing 768 Nvidia A100 GPUs, 30 TB GPU RAM (HBM2), 16 PFlop/s theoretical peak performance (FP64, including tensor cores and host CPUs). HDR InfiniBand interconnect.

IBM Tape library at MPCDF (Photo by Klaus Zilker, MPCDF)

IBM tape library

IBM TS4500 tape library. In total, MPCDF tape libraries provide a capacity of more than 105,000 LTO tape slots and 200 LTO tape drives and store more than 300 Petabytes of long-term archives (HPSS) and backups (IBM Spectrum Protect). The data archives of Max-Planck scientists include experiment and simulation data from many science areas, e.g. from life and materials sciences, astro and fusion research, and also treasures of the World Heritage like sound recordings of endangered languages or images of historic frescoes.

 

National and international projects

National and international projects

The MPCDF, in partnership with different Max Planck Institutes is engaged in several national and international projects and research consortia, such as EU Centers of Excellence, with a focus in high-performance computing, artificial intelligence, data analytics, and data management.

Science use cases

Science use cases

The brochure "High-Performance Computing and Data Science in the Max Planck Society" shows examples of cutting-edge science in the Max Planck Society which were enabled by MPCDF services.

Welcome to the Max Planck Computing and Data Facility (MPCDF)

Mission

The Max Planck Computing and Data Facility (MPCDF, formerly known as RZG) is a cross-institutional competence centre of the Max Planck Society to support computational and data sciences. In close collaboration with domain scientists from the Max Planck Institutes the MPCDF is engaged in the development and optimization of algorithms and applications for high performance computing and data analytics as well as in the design and implementation of solutions for data-intensive projects. The MPCDF operates state-of-the-art supercomputers, several mid-range compute systems and data repositories for various Max Planck institutes, and provides an up-to-date infrastructure for data management including long-term archival.

Operational Information

View the status of MPCDF services and scheduled maintenances.

Help yourself in the MPCDF Selfservice.

In case of problems or questions concering MPCDF services please use one of these communication channels:

News at MPCDF

News

The Workshop on Asynchronous Many-Task Systems and Applications (WAMTA) which will take place from Feb 16-18, 2026 at MPCDF in Garching.

A new issue of the MPCDF computer bulletin Bits&Bytes was released. Read about the HPC system Viper-GPU, various software news, a new LLM inferencing service, updates to the MPCDF gitlab CI infrastructure, a major overhaul of the software ...

A new, dedicated AI system (DAIS) procured by 12 Max Planck Institutes (MPIs) and the MPCDF is operational.

The second phase of the new supercomputer Viper of the Max Planck Society is now open for early user operation.

Job Offers

Upcoming Events

Meet MPCDF: Efficient usage of the tape archive

Feb 5, 2026 03:30 PM - 04:30 PM (Local Time Germany)
online via Zoom

AMD-GPU Workshop, Lectures and Exercises

Apr 21, 2026 09:00 AM (Local Time Germany) - Apr 24, 2026 01:00 PM
online via Zoom

AMD-GPU Workshop, Hackathon

Apr 27, 2026 09:00 AM (Local Time Germany) - Apr 29, 2026 05:00 PM
online via Zoom

Introduction to MPCDF Services

Apr 30, 2026 02:00 PM - 04:30 PM (Local Time Germany)
online via Zoom
Go to Editor View