SuperComputing 2021: oneAPI Acceleration for HPC

As announced by Jeff McVeigh—Intel VP & GM of Data Center XPU Products & Solutions—at SC21, the 2022.1 release of Intel® oneAPI Toolkits will be released in December with more than 900 new features since the 2021 release.

Highlights include :

  • New cross-architecture development capabilities for CPUs and GPUs through the first unified C++/SYCL*/Fortran compiler and data parallel Python. 
  • Advanced analysis features for accelerator modeling and visualizing performance hotspots. 

Additionally, Jeff McVeigh announced the addition of two more oneAPI Centers of Excellence that will accelerate oneAPI ecosystem expansion in HPC and AI domains. These CoEs will deliver strategic code ports, additional hardware support, technologies and services, and developer training. 

Université Old Dominion

Old Dominion University with NASA support will optimize unstructured-grid CFD kernels and develop efficient implementations on Intel® CPUs and Xe architecture GPUs to help solve complex aerodynamic problems.

Institut de technologie informatique de l'Académie des sciences de Chine

Chinese Academy of Sciences Institute of Computing Technology : Intel announces a strategic collaboration with the Institute of Computing Technology of the Chinese Academy of Sciences (CAS-ICT) at the China Super Computing Conference 2021. Intel and CAS-ICT are jointly establishing the first oneAPI Center of Excellence in China to expand oneAPI support on Chinese-developed hardware and using oneAPI to develop full-stack open-source software. This increases CAS-ICT participation in the oneAPI industry initiative. It has contributed to the open-source Data Parallel C++ implementation and are engaged with the oneAPI Technical Advisory Boards.

Source : Intel