High-Performance Computing at TFE
High-performance computing (HPC) is essential for our research in fluid mechanics, thermal science, granular matter and multiphysics applications. At TFE we combine local clusters, faculty-level investments, and the central University of Twente HPC infrastructure. Together, these resources give our researchers the flexibility to move seamlessly from software development to prototyping to full-scale HPC runs.
Computational infrastructure
Central UT HPC cluster
All ET researchers have access to the university’s HPC cluster, which provides:
- CPU and GPU nodes for CFD, DEM, multiphysics and machine learning
- High-capacity storage and robust job scheduling
- Licences for commercial software such as Ansys and Tecplot
New ET compute nodes (2025)
To extend these facilities, ET is adding new compute nodes optimised for MPI scaling:
- 7 × Dell PowerEdge R7725 servers
- 2 × AMD EPYC 9755 per node → 256 cores / 512 threads
- Total: 1,792 cores
- 512 GB memory per node (16 × 32 GB, 6400 MT/s RDIMM)
- 7.68 TB NVMe local storage + 960 GB boot (RAID 1)
- Ultra-low-latency Omni-Path interconnect
Local TFE clusters
TFE also maintains several departmental clusters for development and medium-scale simulations:
- Thermal engineering cluster – 18 nodes, 208 cores (Xeon E5620 & E5-2670 CPUs, 24–64 GB RAM/node)
- Fluid dynamics cluster – 20 nodes, 192 cores (Xeon CPUs) plus a 32-core Bull NovaScale system
- Granular flow cluster – 11 nodes, 396 cores (Xeon E5-2695 v4, 64 GB RAM per node)
Software
TFE has a strong focus on open-source software development, ensuring that methods and tools are transparent, reproducible and widely available.
- Core contributions include open-source codes MercuryDPM, SU2, oomph-lib, pyMOR and APES, as well as in-house solvers EFD Flow and FMG Solver
- Codes are developed and benchmarked on both local and central HPC resources
- Commercial software (e.g. Ansys, Tecplot) are also available via UT licences
Contact
For more information on TFE’s HPC resources, please contact:
Martin Wilens – m.h.g.wilens@utwente.nl