El próximo viernes, 20 de junio, a las 12h., tendrá lugar el seminario CloudHPC - An HPC Cluster for Research, impartido por el Dr. Ruggero Poletto (https://cloudhpc.cloud/).
El evento, organizado por CFD FEA SERVICE SRL, se celebrará en el Campus Río Ebro, edificio I+D, Seminario A (bloque 1, 1ª planta).
CloudHPC is a high-performance computing platform that frees researchers and engineers from running and maintaining their own hardware while they carry out demanding numerical studies. The service supports a wide range of workloads, including computational fluid dynamics (with packages such as OpenFOAM, FDS and Code _Saturne), finite-element and finite-volume analyses (CalculiX, OpenRADIOSS, Code _Aster), acoustics (Fenics), fluid–structure interaction (Solids4Foam, preCICE), machine-learning frameworks (PyTorch, TensorFlow) and post-processing tools like ParaView and SALOME. Users can also launch their own custom codes written in C/C++, Python, R or Fortran.
Underlying this software stack is a cluster built on the latest AMD EPYC Rome, Milan, Genoa and Turin processors alongside Intel Sapphire Rapids CPUs, clocked up to 3.7 GHz. A single run may harness as many as 224 hardware threads (180 physical cores), and an account can queue up to 40 concurrent simulations—with higher limits negotiable with the IT team.
Today the platform is used by manufacturing firms, engineering consultancies and independent specialists who need intensive CFD, FEA/FEM and related analyses but prefer to outsource the compute infrastructure required to deliver them.