A Big Step Forward: Moving Ginkgo to oneAPI

Published: 07/26/2021  

Last Updated: 07/26/2021

[embed]https://w.soundcloud.com/player/?url=https%3A//api.soundcloud.com/tracks/1092640708&color=%23ff5500&auto_play=false&show_artwork=false&visual=false[/embed]


Subscribe Now
Stay in the know on all things CODE. Updates are delivered to your inbox.
Sign Up

 

 

Hartwig Anzt
Research scientist, University of Tennessee

Terry Cojean
Lead developer of Ginkgo software, Karlsruhe Institute of Technology (KIT)

George Silva
Academic programs - program manager, Intel

Ginkgo is a production-ready, sparse linear algebra library used for high-performance computing (HPC) on GPU architectures. It’s now using oneAPI cross-architecture programming to support its foundational design with a high level of performance portability and a focus on software sustainability. Experts Hartwig Anzt and Terry Cojean of Karlsruhe Institute of Technology (KIT) provide their insights on lessons learned from moving CUDA code to other hardware architectures and tools that help smooth that transition. “The oneAPI ecosystem has proven to be a very powerful and useful option for us to actually target different architectures that are all supported by oneAPI.” Listen in. [21:19]

Download Audio  Download Transcript

    

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.