Centralized HPC Infrastructure

Comprehensive Assessment

We continuously conduct a thorough assessment of the research computing needs across various disciplines, identifying areas that heavily rely on CPU, GPU, and storage resources.

Robust HPC Cluster

We will invest in a centralized HPC cluster at our Leibniz Supercomputing Centre (LRZ), which is equipped with high-performance CPUs, GPUs, and ample memory capable of handling demanding computations required by our researchers.

Scalability, Flexibility, and Cloud resources

The HPC infrastructure will accommodate future growth, ensuring scalability and adaptability to evolving research requirements. This also includes the transition from HPC systems to cloud technologies, where we explore additional Hyperscalers resources.

Tiered Structure

In addition to the centralized HPC infrastructure, we will maintain existing compute infrastructure on a school, departmental, and project level to facilitate immediate small to medium-scale computations and to address specific requirements of individual research projects.

Team Research

Contact:

TUM Research Data Hub

By embracing this strategy, we will create a research ecosystem that centralizes HPC resources, optimizes GPU utilization, and establishes robust research data management practices. Together, let us drive groundbreaking discoveries, advance knowledge, and make a lasting impact in our respective fields.

dr-alexander-braun

Dr.-Ing. Alexander Braun

Senior Vice President, CIO

GPU Computing and Optimization

GPU Provisioning

The centralized HPC cluster will include dedicated GPU resources, allowing researchers to harness the power of GPU acceleration for computationally intensive tasks.

GPU Software and Libraries

We will ensure the availability of GPU-specific software, frameworks, and libraries to support researchers in effectively utilizing GPUs, including CUDA, TensorFlow, and other GPU-accelerated technologies.

Optimization Assistance

We will provide guidance and optimization support to researchers, enabling them to leverage GPUs efficiently and maximize their computational capabilities.

High-Performance Storage

Centralized Storage System

We provide a centralized storage solution. This includes NAS and Sync+Share to store and collaborate on data. At the same time, the Data Science Storage (DSS) offered by LRZ is optimized for high-performance data processing, accommodating the storage needs of large-scale research projects.

Data Backup and Replication

Robust backup and replication mechanisms will be established to ensure data integrity and availability, safeguarding research outcomes from potential hardware failures or disasters.

Data Archival and Preservation

We develop archiving and long-term preservation strategies with ISAR, ensuring that research data is securely stored and accessible for future reference or reuse.

Research Data Management

Data Management Planning

Through the TUM Research Data Hub, we assist researchers in developing comprehensive data management plans (DMPs), addressing data collection, storage, sharing, and preservation by regulatory compliance and ethical guidelines.

Data Sharing and Collaboration

We will facilitate secure and controlled data sharing and collaboration platforms within our university and with external partners, promoting interdisciplinary collaboration and fostering research innovation.

Data Governance and Security

To ensure data integrity and compliance, we will establish data governance frameworks that define ownership, access controls, and security measures, protecting sensitive research data.

Research Community Engagement

Training and Workshops

We organize training sessions and workshops to equip researchers with the necessary skills, networks, and knowledge to leverage HPC resources and GPUs and manage their research data effectively.

Digital Skills

Together with TUM Institute for LifeLong Learning, we will offer training and consultation for academic leaders (professors at all career levels: Faculty@TUM) and mid-level academics (CareerDesign@TUM) to equip them with the necessary digital skills, e.g., data science basics and ethics in data management.

Collaboration and Funding Opportunities

We will actively seek partnerships with industry, government agencies, and funding bodies, exploring collaborative research projects, and securing funding to support research initiatives with substantial computing and storage requirements.

Research with personal data, e.g., health data

To enable researchers to work with highly sensitive data such as health data, the TUM will ensure that the IT infrastructure described above will provide the required security measures.

TUM works closely with the Bavarian-wide initiative “Bavarian Cloud for Health Research” (BCHR) to provide secure, scalable, and trusted IT infrastructure. In addition, this will give access to cataloged data within the health sector.

Topics: Research and Innovation

Confluence Migration

Confluence Migration

In response to recent Atlassian licensing changes, we have migrated our Confluence platform to BayernCollab.

Digital.Campus Bayern learning platform available

Digital.Campus Bayern is an initiative of the Bavarian State Ministry for Digital Affairs.
Abbildung zeigt ein Netzwerk aus elektrischen Bahnen

TUM Guideline for the use of AI

The use of AI-supported processes / algorithms and AI-supported tools is steadily increasing. Ki and data protection are not mutually exclusive.

TUM Research Data Hub starts its work

The TUM Research Data Hub is the central point of contact for TUM researchers and partners in all matters of research data management (RDM), following the TUM guidelines for handling research data. Our service portfolio includes consulting, training, networking events, infrastructure solutions, tools, and handouts for all phases of a research cycle.

Grammarly available for all TUM members

Starting May 2023, Grammarly is available to all members of TUM.

Appointment portal online

Appointments to new professorships can now be managed and advertised via TUM’s new appointment portal.

Data science storage available to all scientists

LRZ’s Data Science Storage (DSS) is a novel approach at LRZ to solve the demands and requirements of data-intensive science.