Optimize Edge Compute Performance by Migrating CUDA to DPC++

Thorsten Moeller, Intel
Intel webinars

Optimize Edge Compute Performance by Migrating CUDA to DPC++

Edge Computing is synonymous with faster, more stable, real-time services that avoid bandwidth constraints. But how do industries accomplish this? Join this session to learn how customers in the Healthcare and Life Sciences industries did it … by using oneAPI to create performant apps and offload them to the GPU.

Edge computing is a must for real-time data processing. Manufacturing plants, connected and autonomous vehicles, financial services, AR/VR, and virtual assistants are just some of the industry use cases with enormous potential.

This webinar focuses on two others: Healthcare and Life Sciences. In particular, it looks at how these industries have used oneAPI to develop compute-intense edge applications that were subsequently (and successfully) offloaded to the GPU.

Join Intel Sr. Product Manager Thorsten Moeller for a use case walk-though, where he’ll cover:

  • How to migrate CUDA code to Data Parallel C++ code using the Intel® DPC++ Compatibility Tool
  • How to use the Intel® oneAPI DPC++/C++ Compiler to increase application performance and offload compute-intensive workloads to the GPU
  • How to use Intel® VTune™ Profiler to tune medical application performance