2022-11-12, 16:30–16:45 (US/Central), Room
Heterogeneous computing platforms are continuously evolving as hardware vendors (Intel, AMD, NVIDIA etc.) are increasingly focusing on architectural innovations
targeting scientific computing applications. Consequently, GPU computing is no longer confined to a single vendor hardware and programming model. SYCL is becominga de facto standard for vendor agnostic heterogeneous computing. Upgrading CUDA code to standard C++ with SYCL makes the applications portable across a range of existing and evolving accelerators including Nvidia GPUs. In this talk, we will review the state-of-the-art in heterogeneous computing and the current status of vendor agnostic tools.
Chekuri works us for Intel where he leads the migration of scientific applications written in CUDA to SYCL. In his past role at IBM, he developed technical roadmaps and strategies including GPU utilization for machine learning software named Bayesian Optimization Accelerator. He worked with several clients and universities across the globe evangelizing AI and designing machine learning solutions by integrating IBM products with open-source software. Prior to his work at IBM, he was an applied researcher in AI and High-Performance Computing collaborating with the U.S. Department of Energy National Labs, NASA and universities. As part of the scientific computing team in ExxonMobil, he accelerated scientific software using MPI and CUDA, and wrote C++ for seismic wave inversion. He has a PhD in Computer Science and Engineering from the University of South Carolina.