Lichtenberg High Performance Computer of TU Darmstadt
At TU Darmstadt, the Lichtenberg High Performance Computer provides computing resources for scientific research for scientists of different Hessian universities.
The multifaceted architecture of the high performance computer allows for flexible and efficient scientific computing, especially for computationally intensive applications. It supports the efficient execution of a variety of existing programs, as well as the development of new parallel programs.
The experts of the “Hessischen Kompetenzzentrum für Hochleistungsrechnen” (Hessian Competence Center for High Performance Computing) will provide for consultancy and scientific support on using the cluster efficiently. The HKHLR also hosts various (regular) workshops on scientific HPC.
The high performance computer has been established in two phases. The first phase has already been installed in 2013, the second phase was installed 2015. In total, the two phases comprise of about 1400 nodes, 35 times the computing power of the old cluster (“HHLR”). In its current configuration, the Lichtenberg cluster has a theoretical peak performance of approximately 1 PFLOP/s.
Funding for the acquisition and and operation is provided by TU Darmstadt, the state of Hesse and the German Federal Government.
The regular consultation is open to all (potential) users, with all kinds of questions about project proposals for or resources of the Lichtenberg supercomputer. go
For the (potential) users of the Lichtenberg supercomputer, a monthly introduction will be held in Darmstadt. Subjects are the available hardware and the general use of the (batch) system. go
Deep Learning, Tools and Fortran Modernization
The 11th installment of the HiPerCH workshop series takes place at Technical University Darmstadt on September 23-27, 2019. The workshop focuses on Deep Learning on high performance computing clusters, and related tools and languages. In addition, we hold a course on Fortran modernization in cooperation with NAG.
The HiPerCH workshops are conducted by the Competence Center for High Performance Computing in Hessen (HKHLR). They are targeted at students and scientists from Hessen, Mainz, and Kaiserslautern with interest in programming modern HPC hardware. go
Unterstützung der KI-Forschung durch Erweiterungen der Programmiersprache Python und Container
Forschung zu künstlicher Intelligenz (KI), insbesondere zu maschinellem Lernen (ML), ist in der computergestützten Forschung der letzten Jahre das zentrale Thema. Um die Wissenschaftler_innen bei der Deep Learning- und ML-Forschung auf dem Lichtenberg Cluster zu unterstützen und für seine Nutzung zu gewinnen, wurden alle im Modulsystem vorhandenen Python Installationen ergänzt. weiter
For a major storage migration inside the Lichtenberg Cluster, a two-day maintenance has been scheduled from Tuesday, 23rd to Wednesday, 24th of July 2019. Neither the compute nodes nor the login nodes will be available. go