This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| CINECA MARCONI | |
|---|---|
| Name | CINECA MARCONI |
| Caption | Supercomputing system installed at CINECA |
CINECA MARCONI is a high-performance computing system hosted at a major Italian research consortium that supports scientific computing, simulation, and data analysis. It serves academic, industrial, and governmental research communities by providing computational resources for fields spanning astrophysics, climate science, molecular dynamics, and artificial intelligence. The system integrates advanced processors, accelerator technologies, and high-speed storage to deliver large-scale parallel computing for international projects and national initiatives.
CINECA MARCONI operates within a national research infrastructure alongside PRACE projects, EuroHPC initiatives, and collaborations with institutions such as INFN, CNR, ENEA, Sapienza University of Rome, Politecnico di Milano, University of Bologna, Scuola Normale Superiore, University of Pisa, University of Padua, University of Turin, University of Milan, University of Naples Federico II, University of Rome Tor Vergata, and University of Trento. It supports workloads from research programs linked to Horizon 2020, Horizon Europe, ESFRI, CERN, ESA, ESO, ICGEB, IFReC, EMBL, EUMETSAT, Copernicus, IPCC, and WMO. The facility provides access to users through peer-reviewed allocations, competitive calls, and industrial partnerships involving Leonardo S.p.A., ENI, Thales Alenia Space, Stellantis, and Fincantieri.
The architecture integrates multicore CPUs, GPUs, and interconnect fabrics from vendors including Intel Corporation, AMD, NVIDIA, IBM, Cray Inc., HPE, Dell EMC, Lenovo, and Fujitsu. Compute nodes combine architectures such as x86-64, ARM architecture, and accelerator modules comparable to NVIDIA Tesla, NVIDIA A100, AMD Instinct, and Intel Xeon Phi concepts. The network topology leverages technologies like InfiniBand, Omni-Path, and high-speed Ethernet standards, and storage layers use parallel filesystems influenced by Lustre, GPFS/IBM Spectrum Scale, and BeeGFS. Cooling and power systems reference solutions from Schneider Electric, Siemens, ABB Group, and designs paralleling installations at Oak Ridge National Laboratory, Lawrence Livermore National Laboratory, Argonne National Laboratory, National Supercomputing Centre (Norway), and Barcelona Supercomputing Center.
Performance is assessed with benchmarks such as the High Performance Linpack, HPCG, STREAM benchmark, IO500, SPEC CPU, and application-level metrics drawn from GROMACS, LAMMPS, Quantum ESPRESSO, VASP, NAMD, OpenFOAM, WRF, COSMO, and ECHAM. Comparative rankings relate to entries on the TOP500, Green500, and national ranking lists maintained by organizations like CINECA and GARR. Optimization strategies reference compiler suites from GNU Compiler Collection, Intel Parallel Studio, PGI/NVIDIA HPC SDK, and libraries such as OpenMPI, MPICH, CUDA, ROCm, OpenACC, OpenMP, MKL, BLAS, LAPACK, and FFTW.
The software stack supports operating systems and management tools aligned with Linux kernel distributions, cluster managers like Slurm Workload Manager, PBS Professional, Univa Grid Engine, and provisioning systems such as xCAT and OpenStack for cloud-like services. Development environments include languages and ecosystems like Fortran, C++, C, Python (programming language), R (programming language), Julia (programming language), and frameworks such as TensorFlow, PyTorch, Keras, Scikit-learn, HDF5, NetCDF, Paraview, VisIt, and Jupyter Notebook. Middleware and data management draw on Globus, iRODS, Apache Hadoop, Apache Spark, and provenance tools used by EOSC projects.
Use cases cover computational fluid dynamics linked to ANSYS, OpenFOAM, and SU2; climate and Earth system modeling with ECMWF, CMIP, EC-Earth, and IPSL-CM5A; astrophysics simulations using ENZO, GADGET, and RAMSES; bioinformatics and structural biology workflows involving BLAST, BWA, GATK, AlphaFold, and Rosetta; materials science with VASP, Quantum ESPRESSO, and ABINIT; and machine learning research for applications in autonomous vehicles, medical imaging, drug discovery, and financial modeling. Industrial collaborations apply to sectors represented by Fiat Chrysler Automobiles, Pirelli, ENEL, Saipem, Edison (company), and TIM (company).
The system evolved within a trajectory connecting European supercomputing milestones including installations at CINECA, participation in PRACE infrastructures, alignment with EuroHPC Joint Undertaking, and procurement cycles reflecting trends established by systems at Jülich Supercomputing Centre, TGCC, Leibniz Supercomputing Centre, and Met Office. Development involved collaborations with vendors such as Hewlett Packard Enterprise, Lenovo, Cray Inc., and NVIDIA. Research programs and funding rounds tied to MIUR, Italian Ministry of Education, Universities and Research, and the European Commission influenced upgrades and decommissioning schedules akin to patterns at Sequoia (supercomputer), Fugaku, and Piz Daint.
Procurement processes engaged public tenders and framework agreements similar to those managed by Consip, European Investment Bank, Italian National Recovery and Resilience Plan, Horizon 2020, Horizon Europe, and regional funds from ERDF. Funding partners included national ministries, academic consortia such as CINECA members, private industry partners like Leonardo S.p.A. and ENI, and European bodies like European Commission programs. Contractual and compliance frameworks referenced procurement law precedents and transparency practices comparable to EU procurement law.
The system contributed to high-impact outcomes published with partners including Nature (journal), Science (journal), Physical Review Letters, The Astrophysical Journal, Journal of Chemical Physics, Geophysical Research Letters, and collaborations with laboratories such as CERN, INFN Laboratori Nazionali del Gran Sasso, ESA/ESRIN, INEA, EIT Digital, ELIXIR, EMBL-EBI, Human Brain Project, and networks like EUDAT. Educational and training activities connected to PRACE Summer of HPC, RISE, EuroHPC Competence Centers, and university curricula at Politecnico di Torino, University of Bologna, University of Milan-Bicocca, Scuola Superiore Sant'Anna, and Gran Sasso Science Institute. The platform supported industrial HPC adoption pathways reflected in case studies with Leonardo S.p.A., ENI, Saipem, and startups incubated by Polihub and Luiss Enlabs.
Category:Supercomputing