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.
| Cray XC30 | |
|---|---|
| Name | Cray XC30 |
| Manufacturer | Cray Inc. |
| Family | XC |
| Released | 2012 |
| Os | CLE, SUSE Linux Enterprise Server |
| Cpu | Intel Xeon |
| Gpu | NVIDIA Tesla |
Cray XC30 is a high-performance computing system introduced by Cray Inc. designed for scientific simulation, data analysis, and large-scale modeling. It integrates compute blades, Aries interconnect networks, and scalable software stacks to serve research institutions, national laboratories, and industry partners. The system was deployed in academic, governmental, and commercial centers worldwide and influenced subsequent supercomputer architectures.
The XC30 combined commodity Intel Xeon processors with Cray's proprietary Aries interconnect to target workloads in climate science, computational chemistry, and astrophysics. Early adopters included facilities associated with the NASA, U.S. Department of Energy laboratories, and university consortia such as University of Cambridge, Stanford University, and University of Tokyo. The platform competed with systems from IBM, Hewlett-Packard, and Fujitsu in Top500 listings including installations tied to the ECMWF, Los Alamos National Laboratory, and other centers.
The system architecture built around compute blades populated with dual-socket Intel Xeon processors and optional accelerator blades featuring NVIDIA Tesla GPUs or Intel Xeon Phi co-processors. A key element was the Aries network chip providing low-latency, high-bandwidth 3D torus or dragonfly-like topologies interconnecting cabinets; Aries traced design lineage from earlier Cray systems such as predecessors developed after Seymour Cray's era and contemporaneous with projects at Sandia National Laboratories and Los Alamos National Laboratory. The XC30 used Cray's hardware management and scheduler integration compatible with resource managers like PBS Professional, SLURM, and batch systems used at Oak Ridge National Laboratory and Argonne National Laboratory.
XC30 installations achieved high positions on the TOP500 list when configured at scale, outperforming many contemporaneous clusters on workloads like LINPACK, HPCG, and application kernels used by NOAA, Met Office, and research consortia at ETH Zurich. Benchmark suites run on XC30 systems included scalable tests for weather prediction codes such as the WRF, ocean models used by Scripps Institution of Oceanography, and quantum chemistry packages used at Lawrence Berkeley National Laboratory. Performance tuning exploited vectorization for Intel instruction sets and accelerator offload for CUDA-enabled codes developed at institutions like Los Alamos and CERN.
The XC30 shipped with Cray Linux Environment (CLE), integrating the SUSE Linux Enterprise Server distribution and system software libraries tailored for HPC. Programming models supported included MPI implementations used in projects at Max Planck Society, Princeton University, and Imperial College London, as well as shared-memory paradigms like OpenMP used by research groups at MIT and Caltech. For accelerators, the environment supported CUDA for NVIDIA GPUs and frameworks used in collaborations with Lawrence Livermore National Laboratory, while performance tools included profilers and debuggers similar to those used at Argonne and NERSC.
Notable XC30 deployments included systems installed at national facilities such as Los Alamos National Laboratory, Austrian Institute of Technology, and the United Kingdom Meteorological Office where they supported operational forecasting and research. Academic installations at institutions like ETH Zurich, University of Tokyo, and University of Copenhagen enabled studies in climate modeling, materials science, and genomics. Commercial and government clients ranged from research divisions of Airbus and BP to defense science organizations analogous to DARPA-funded collaborations, often interfacing with national grid initiatives and regional supercomputing centers.
Physical design emphasized modular cabinets and liquid-cooling options to manage power density for racks populated with Intel Xeon CPUs and NVIDIA Tesla accelerators. Data centers hosting XC30 systems incorporated chilled-water infrastructure, raised-floor arrangements seen in installations at NERSC and Met Office, and power distribution strategies aligned with standards adopted by DOE facilities. Energy efficiency metrics and PUE targets were important for operators such as Oak Ridge National Laboratory and European centers participating in PRACE initiatives.
Launched in the early 2010s, the XC30 represented an evolution of Cray's product line following systems like those produced in the eras of partnerships involving Seymour Cray's legacy and corporate developments leading to acquisitions and collaborations with companies such as Silicon Graphics International and HPE-era competitors. Development teams collaborated with national labs including Argonne National Laboratory and user communities at European Centre for Medium-Range Weather Forecasts to refine interconnect technologies and software stacks. Subsequent generations incorporated lessons from XC30 deployments into designs that influenced later Cray and industry systems used at CERN, NASA Ames Research Center, and major computational science consortia.