A developer workflow for AI Tools uses techniques based on cross-vendor and cross-architecture standards.

The presenters share several use cases that include code tutorials and demos. These presentations represent the latest and most interesting aspects of oneAPI programming:

1. oneAPI-based GPU porting of a thermal solver and seismic modeler that use the Intel® oneAPI DPC++/C++ Compiler to create vendor-agnostic SYCL* code (from the Khronos Group*)
2. Intel® Fortran Compiler offload to GPUs using:
   • Auto offload through native language features and compilation techniques
   • OpenMP* on GRILLIX/PARALLAX, a plasma turbulence and transport simulator

AI Tools and SYCL provide a standards-based programming model that enables code portability across computing architectures (including CPUs, GPUs, and FPGAs) and across vendors. This standards-based approach increases developer productivity that avoids the vendor lock-in inherent with proprietary languages.

Speakers

Rafael Lago has earned a PhD in computer science. He has a strong academic background and worked in research institutes in Brazil, Canada, France, and Germany. Rafael specializes in numerical linear algebra and message passing interface (MPI), and applies his expertise to seismology, geophysics, plasma physics, and applied mathematics applications. In January 2022, Rafael joined Intel as a technical consulting engineer (TCE) in the Europe, Middle East, and Africa (EMEA) team.

Rupak Roy is a software TCE at Intel. He has expertise in analyzer tools for Intel oneAPI, Intel compilers, heterogeneous programming, and high-performance computing (HPC). Rupak has also worked with the Intel® Video Processing Library (Intel® VPL).

Shiquan Su is a software TCE in the Developer Tools Customer Engineering department. His degree is in theoretical physics, and his career focuses on HPC. Shiquan supports the Intel Fortran Compiler, OpenMP*, and other products in the oneAPI programming model.

James Tullos is a senior TCE with Intel® Software Development Products. He enables developers to use tools for HPC and parallel programming.

Yantao Zhang is a software TCE with Intel. He supports the Intel® oneAPI DPC++ Compiler and Intel® DPC++ Compatibility Tool. His interests include GPU programming and architecture, and computational science. Yantao has a PhD in computational mechanics from the structural engineering division at the University of California San Diego.