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.
| Frontera (supercomputer) | |
|---|---|
| Name | Frontera |
| Manufacturer | Dell Technologies |
| Operator | Texas Advanced Computing Center |
| Location | University of Texas at Austin |
| Cost | $60 million |
| Power | 7.5 MW |
| Speed | 23.5 PFLOPS (peak) |
| Memory | 1.7 PB |
| Storage | 20 PB Lustre |
| Os | CentOS |
Frontera (supercomputer) is a high-performance computing system deployed at the University of Texas at Austin and operated by the Texas Advanced Computing Center. Funded by the National Science Foundation and collaborations with Dell Technologies and Hewlett Packard Enterprise, it served as a national resource for scientific computation, ranked among the world's fastest systems and supporting research across disciplines such as climate science, astronomy, bioinformatics, materials science, and artificial intelligence.
Frontera provided researchers at institutions including Argonne National Laboratory, Lawrence Berkeley National Laboratory, Oak Ridge National Laboratory, Los Alamos National Laboratory, and the National Center for Supercomputing Applications with access to high-performance resources for projects funded by the Department of Energy, National Institutes of Health, and the National Aeronautics and Space Administration. As part of initiatives related to the Exascale Computing Project and national cyberinfrastructure strategies, Frontera supported workflows spanning from large-scale simulations used by teams at NASA and European Southern Observatory to data analytics performed by groups at Google Research and IBM Research.
Constructed using servers from Dell EMC and interconnect components from Intel, Frontera’s architecture combined 8,008 nodes of Intel Xeon Platinum processors with a high-speed Mellanox (now part of NVIDIA) InfiniBand fabric. The system integrated technologies found in platforms by HPE Cray and incorporated dense memory configurations similar to those used in clusters at CERN and Fermilab. Storage subsystems used Lustre parallel file systems and enterprise SSDs used by enterprises like Amazon Web Services for metadata acceleration. Frontera’s design reflected influences from earlier systems such as Blue Waters, Titan (supercomputer), and Sierra (supercomputer).
Frontera achieved peak performance measured in the petaflops range and participated in benchmarks including LINPACK and domain-specific tests used by the TOP500 and Green500 lists. Comparative analyses referenced performance characteristics of systems like Summit (supercomputer), Fugaku, and Tianhe-2. Benchmark suites used by research teams included HPL, HPCG, and applications from the SPEC and NAS Parallel Benchmarks families; these were leveraged by projects collaborating with Los Alamos National Laboratory and Sandia National Laboratories to validate scientific codes.
Frontera supported operating environments based on CentOS and common cluster management tools influenced by software stacks from OpenStack and SLURM Workload Manager. Users compiled codes with toolchains from Intel Corporation and compilers like GNU Compiler Collection alongside performance libraries such as Intel MKL, OpenBLAS, and FFTW. Scientific packages and frameworks deployed on Frontera included TensorFlow, PyTorch, LAMMPS, GROMACS, NAMD, and community codes used at Princeton University and Massachusetts Institute of Technology. Middleware for provenance, job scheduling, and data movement drew from projects at Argonne National Laboratory and National Center for Atmospheric Research.
Researchers employed Frontera for large-scale climate modeling with collaborations involving NOAA and IPCC contributors, cosmological simulations akin to work at Space Telescope Science Institute and European Space Agency, and genomic analyses comparable to efforts at Broad Institute and Wellcome Sanger Institute. Computational chemistry and materials discovery projects interfaced with databases and tools maintained by Materials Project and Lawrence Berkeley National Laboratory scientists. Machine learning research leveraged deep learning frameworks to accelerate applications relevant to teams at OpenAI and DeepMind while supporting visualization and data analysis pipelines used by Princeton Plasma Physics Laboratory.
Frontera operated within a datacenter environment using power distribution and cooling strategies similar to those developed at Argonne National Laboratory and Lawrence Livermore National Laboratory, drawing up to multiple megawatts and employing chilled-water and hot-aisle containment schemes like installations at Oak Ridge National Laboratory. Facility management used monitoring and energy-efficiency practices evaluated by the Green Grid and measured on lists such as Green500. Backup power and electrical infrastructure were coordinated with campus utilities at University of Texas at Austin and complied with standards promoted by IEEE.
Announced after awards from the National Science Foundation, Frontera was installed and commissioned by teams including staff from Texas Advanced Computing Center and contractors such as Dell Technologies and Intel. Its deployment followed a lineage of NSF-funded systems including Stampede2 and preceded newer efforts in exascale computing coordinated by the Exascale Computing Project and national laboratories like Oak Ridge National Laboratory. Over its lifecycle, Frontera received software and firmware updates and informed design choices for successors at institutions such as Argonne National Laboratory and projects tied to DOE Office of Science priorities.