StroemungsRaum - High performance computation for realistic applications
Novel Exascale-ready methods with heterogeneous hardware components for Computational Fluid Dynamics Simulations in the Stroemungsraum platform.
About StroemungsRaum Contact Us
The StroemungsRaum Project
Combining Efficient, Robust and Verified HPC-Software
Available HPC methods need to be executed ressource-efficiently and need to exploit the performance of future Exascale computers. Integration of these techniques into application software accessible by non HPC-Experts is crucial to deliver these huge computational benefits to relalistic applications. The open-source software FEATFLOW, developed at TU Dortmund University which is a central component of the StroemungsRaum that is successfully used by the industriual partner of the project IANUS for years is the integration platform for our HPC techniques.
High Performance Numerics
Exascale Ready
Industrial Applications
Accessibility
Our Expertise
StroemungsRaum High-Performance Computing
Sustainable Performance Engineering
Support of future and current hardware developments by using performance model aided algorithmic co-design.
Read MoreScalable Multilevel Methods
Increase the scalability of existing methods by tailoring them to new parallel architectures and by novel decomposition schemes.
Read MoreFROSch - Domain Decomposition
Fast and Robust Overlapping Schwarz (FROSch) domain decomposition can extend the scalability of fluid solvers to exascale levels.
Read MoreParallel In Time Integration
Coupled space and time parallelization on supercomputers, to develop new parallel-in-time integrators increase scalability.
Read MoreSupercomputers & Acceleration Hardware
Performance studies of the fluid solvers on high-ranking clusters, GPUs and new architectures provide insights to algorithmic design
Read MoreStroemungsraum for Industrial Applications
In the StroemungsRaum framework the newly developed software can be tested in complex industrial applications.
Read MoreSustainable Performance Engineering, State-Of-The-Art code analysis and hardware optimization
Friedrich-Alexander-Universität Erlangen-Nürnberg
High-Performance Computing, Parallel Preconditioning, Parallel-In-Time Methods, Exascale Supercomputing
Forschungszentrum Jülich
Highly parallel algebraic Multi-Level-Methods for PDEs, non-linear Domain Decomposition, Multi-Level-FROSch Solver
TU Bergakademie Freiberg
FEATFLOW package with massively parallel, hardware-oriented numerics, basis for next gen schemes like Parallel-In-Time methods
TU Dortmund
Industrial application of latest scientific methods, Stroemungsraum framework offers Simulation-as-a-Service
IANUS Simulation
Scalable parallel solvers for PDEs focusing on non linear Domain Decomposition and Multi-Level-Methods
Universität zu Köln
Project Gallery
Template designed by HTML Codex
Template distributed by: ThemeWagon