LLMpediaThe first transparent, open encyclopedia generated by LLMs

MareNostrum (supercomputer)

Note: This article was automatically generated by a large language model (LLM) from purely parametric knowledge (no retrieval). It may contain inaccuracies or hallucinations. This encyclopedia is part of a research project currently under review.
Article Genealogy
Parent: Blue Gene Hop 5 terminal

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.

MareNostrum (supercomputer)
NameMareNostrum
CaptionMareNostrum supercomputer at the Barcelona Supercomputing Center
ManufacturerIBM
OperatorBarcelona Supercomputing Center
LocationBarcelona, Spain
Date2004–present

MareNostrum (supercomputer) is a series of high-performance computing systems hosted at the Barcelona Supercomputing Center in Barcelona, Spain. The systems bearing the MareNostrum name have supported computational science across disciplines, enabling projects associated with institutions such as the European Space Agency, CERN, European Commission, Max Planck Society, and the Spanish National Research Council. MareNostrum installations have been noted for colocating within a historic setting at the Torre Girona chapel and for participating in pan-European initiatives like the PRACE infrastructure, the European Strategy Forum on Research Infrastructures, and collaborations with National Center for Supercomputing Applications partners.

Overview

MareNostrum has functioned as a national flagship for Spanish and European high-performance computing. The project links the Barcelona Supercomputing Center with Spanish ministries and European funding mechanisms, including the Seventh Framework Programme and Horizon 2020. MareNostrum nodes have been used by researchers from universities such as the University of Barcelona, Pompeu Fabra University, University of Zaragoza, and by laboratories like Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and Argonne National Laboratory. MareNostrum systems have been referenced alongside machines like IBM Roadrunner, Cray XT5, Fujitsu K, and Sunway TaihuLight in rankings produced by the TOP500 list and workload analyses by SPEC.

History and development

The MareNostrum lineage began with installations in 2004 when the Ministerio de Ciencia e Innovación supported the deployment at the Torre Girona site. Subsequent generations paralleled European investments in exascale roadmap planning championed by entities including the European Union and research consortia like Gauss Centre for Supercomputing. Upgrades reflected shifts from vector processors to multi-core x86 clusters and, later, heterogeneous accelerators similar to those adopted by Oak Ridge National Laboratory and Texas Advanced Computing Center. Institutional partners over time included the Spanish Ministry of Science, CSIC (the Spanish National Research Council), and regional bodies such as the Catalan Government.

Architecture and hardware

MareNostrum generations have combined commodity and proprietary components. Early configurations used IBM systems while later installations incorporated Intel Xeon processors, NVIDIA GPUs, and high-performance interconnects comparable to InfiniBand fabrics used at Lawrence Berkeley National Laboratory. Storage architectures evolved from network-attached storage to parallel file systems analogous to Lustre and object stores resembling solutions used at EMBL-EBI. Cooling approaches varied, with rack-based air cooling in initial deployments and experimentation with liquid cooling and energy reuse in later models akin to techniques explored at ETH Zurich and National Renewable Energy Laboratory.

Performance and benchmarks

MareNostrum entries have appeared on the TOP500 list, with peak LINPACK numbers reflecting generational changes comparable to contemporaries such as Blue Gene/P, Frontera (supercomputer), and Piz Daint. Benchmarking campaigns have included LINPACK, HPCG, and domain-specific tests used by groups at NASA Ames Research Center and European Centre for Medium-Range Weather Forecasts. Performance tuning involved compiler toolchains from vendors like Intel Corporation and NVIDIA, and middleware stacks similar to OpenMPI and Slurm Workload Manager used at facilities including CSCS.

Notable installations and upgrades

Significant milestones include the original 2004 installation in the Torre Girona chapel, a mid-2010s upgrade to a cluster emphasizing energy efficiency, and a late-2010s installation that integrated accelerators for machine learning workloads, reflecting trends seen at Google and Microsoft Research. Partnerships for specific upgrades involved corporations such as IBM and Dell EMC and collaborations with European centers like BSC peers in the PRACE network. Each upgrade aimed to support projects funded by entities including the European Research Council and national research programmes.

Applications and research uses

MareNostrum systems have been applied across computational fluid dynamics for groups at Imperial College London, climate modeling for teams at Met Office and Instituto de Ciencias del Mar, bioinformatics projects affiliated with European Bioinformatics Institute, and cosmology simulations used by researchers linked to Institute of Space Sciences (Spain). Use cases extended to materials science collaborations with Max Planck Institutes, seismic modeling with Spanish Geological Survey, and epidemiological simulations akin to work conducted by Johns Hopkins University during public health crises. The platform has supported software frameworks such as GROMACS, OpenFOAM, and TensorFlow.

Management, funding, and facilities

Operational management resides at the Barcelona Supercomputing Center under governance involving the Spanish Ministry of Science and Innovation and academic partners like the Polytechnic University of Catalonia. Funding has combined national allocations, European grants, and industry contributions from companies such as Intel and NVIDIA. Facility aspects include campus security protocols comparable to those at Argonne, energy contracts influenced by regional utilities, and outreach programs that coordinate with educational institutions like IESE Business School and the University of Barcelona for training and workforce development.

Legacy and impact on supercomputing

MareNostrum's legacy is visible in European HPC capacity building, contributing to initiatives led by the European High Performance Computing Joint Undertaking and influencing designs discussed at conferences such as SC (Supercomputing Conference) and ISC High Performance. The series fostered expertise that fed into national centers like BSC and into collaborations with international laboratories such as CERN and Los Alamos National Laboratory. MareNostrum helped demonstrate integration of heritage architecture into modern research workflows, informing procurement strategies at institutions including CNRS and shaping educational curricula at universities like University of Barcelona and Pompeu Fabra University.

Category:Supercomputers Category:Science and technology in Spain