HPC

Since 20:10 on January 30, the parallel Lustre file system ($FASTTMP) on the Fritz cluster is unavailable due to repeated server crashes. The issue is being investigated. Update 2025-01-31 17:00: The issue turned out to be a faulty component in the Infiniband network of Fritz, the Lustre servers ...

Category: Allgemein, HPC, Systems

Since Wednesday (2025-01-22) evening, some nodes in TinyGPU and TinyFAT seem to have problems accessing NFS directories. This will lead to issues like Jobs mysteriously hanging, or interactive jobs starting but never opening a shell prompt. Update 2025-01-23 09:30: This issue should now be resolv...

Category: Allgemein, HPC, Systems

When Europe's most powerful supercomputer JUPITER starts operating in May 2025, Germany's best generative AI model will be determined in a competition. Last Friday, the Federal Ministry of Education and Research (BMBF) opened the “Gauss AI Compute Competition,” which is conducted by the Gauss Center...

Category: Allgemein, HPC

New year, new opportunities to learn: we are happy to announce the NHR@FAU course program for spring 2025. Unless otherwise indicated, all courses include hands-on components. Hybrid Programming in HPC – MPI+X, Three-day hybrid tutorial at HLRS, Stuttgart, January 21-23, 2025. Registration clos...

Category: Allgemein, HPC, Teaching

We are happy to announce a bugfix release of LIKWID (5.4.1). This is what got fixed: Fixes linking errors due to missing bstrlib.h Fix for likwid-bench kernel stream_mem Fix builds with CUDA for versions 11.2 to 12.6 Fix sysfeatures with ACCESSMODE=perf_event Add AMD Zen1 to Zen...

Category: HPC, Tools

Today at about 10:45 a.m., a large-area power outage in Erlangen has brought all our clusters down. All running jobs have been terminated. Some frontends have been switched off manually to lower the load on the uninterruptible power supply infrastructure. We are working to get the systems running...

Category: Allgemein, HPC, Systems