SYCL is a powerful way to enable multi-vendor support for high performance libraries, languages, and packages, while still allowing the originally desired programmer productivity and performance. While many BoFs/Presentations will focus on the end results of a port, this BoF is meant to share lessons learned in porting a diverse array of previously vendor-specific implementations to SYCL, how they enforced numerical reproducibility, and then added flexible vectorization for portable performance.

In this talk we will explore the various options and possible formulations for writing 2D/3D stencil loops using SYCL, then proceed to analyze and contrast their performance on a variety of Intel® Xeon® CPUs and Intel® Iris® Xe GPU s, showing performance improvements of 1.5-2x . By utilizing the OPS framework, we evaluate a number of proxy and production applications, including CloverLeaf and OpenSBLI, running on a single device, and scaling to multiple devices using MPI+SYCL.