July 6, 2023 – Lawrence Livermore National Laboratory (LLNL) made a groundbreaking announcement today regarding the commencement of the assembly of El Capitan, a supercomputer that is poised to revolutionize computational capabilities. After years of dedicated efforts, LLNL is now on track to unveil the world’s first 2 ExaFLOPS supercomputer next year.
With the installation of El Capitan underway, LLNL took to Twitter to share their excitement and the significant progress made thus far. The road ahead remains challenging, but the realization of their long-standing vision is within reach, igniting a sense of exhilaration throughout the laboratory.
El Capitan, like its predecessors Frontier and Aurora, is built upon the HPE Shasta supercomputer architecture, which signifies its construction by HP Enterprise. Leveraging the cutting-edge capabilities of AMD’s highly anticipated Instinct MI300, El Capitan’s key components will feature a 5nm manufacturing process. The CPU module will consist of 24 Zen 4 architecture cores, complemented by a GPU module featuring the CDNA 3 architecture. These components will be accompanied by a robust 128GB HBM3 memory capacity, incorporating the next generation of Infinity Cache and the utilization of the fourth-generation Infinity architecture. Furthermore, El Capitan will support the CXL 3.0 ecosystem, enhancing its compatibility and interoperability.
The assembly of El Capitan symbolizes a significant milestone in the realm of supercomputing, promising unprecedented computational power and performance. LLNL’s dedication and perseverance have propelled them to the forefront of scientific research and technological innovation. As the journey towards the completion of El Capitan continues, the scientific community eagerly awaits the dawn of a new era in high-performance computing, poised to unlock vast realms of knowledge and tackle complex challenges across various domains.