PI Name: Salman Habib
PI Institution: argonne-national-laboratory
Collaborating ANL Division: high-energy-physics-hep

Description: High-energy physics (HEP) experiments have developed millions of lines of code optimized to run on traditional x86 systems. However, leadership-class computing facilities and traditional data centers use new accelerator architectures, such as GPUs. HEP experiments are now faced with the untenable prospect of rewriting millions of lines of code. This task is made more challenging by the architecture specific languages and APIs promoted by the different vendors. The Portable Parallelization Strategies team of the HEP Center for Computational Excellence is investigating Kokkos, SYCL, Alpaka, etc. as potential portability solutions using representative use cases from DUNE, ATLAS, and CMS experiments.

Testbed: Arcticus and other Intel PVC resources; AMD, NVIDIA GPUs as available

© 2020 Joint Laboratory for System Evaluation