Topic: Hybrid/Heterogeneous

Supercomputers broke the exascale barrier, marking a new era in processing power, but the energy consumption of such machines cannot run rampant.

UCLA's Institute for Pure & Applied Mathematics hosted LLNL's Erik Draeger for a talk about the challenges and possibilities of exascale computing.

Combining specialized software tools with heterogeneous HPC hardware requires an intelligent workflow performance optimization strategy.

LLNL is participating in the 34th annual Supercomputing Conference (SC22), which will be held both virtually and in Dallas on November 13–18, 2022.

LLNL participates in the CMD-IT/ACM Richard Tapia Celebration of Diversity in Computing Conference (Tapia2022) on September 7–10.

El Capitan will have a peak performance of more than 2 exaflops—roughly 16 times faster on average than the Sierra system—and is projected to be several times more energy efficient than Sierra.

A Livermore-developed programming approach helps software to run on different platforms without major disruption to the source code.

FGPU provides code examples that port FORTRAN codes to run on IBM OpenPOWER platforms like LLNL's Sierra supercomputer.

Umpire is a resource management library that allows the discovery, provision, and management of memory on next-generation architectures.

Highlights include debris and shrapnel modeling at NIF, scalable algorithms for complex engineering systems, magnetic fusion simulation, and data placement optimization on GPUs.

Highlights include the latest work with RAJA, the Exascale Computing Project, algebraic multigrid preconditioners, and OpenMP.

Highlights include recent LDRD projects, Livermore Tomography Tools, our work with the open-source software community, fault recovery, and CEED.

Highlights include the HYPRE library, recent data science efforts, the IDEALS project, and the latest on the Exascale Computing Project.

Livermore computer scientists have helped create a flexible framework that aids programmers in creating source code that can be used effectively on multiple hardware architectures.