The Argonne Leadership Computing Facility (ALCF) in partnership with Intel will host the ALCF Aurora Learning Paths learning seriesto explore use of oneAPI and Data Parallel C++ (DPC++), Intel’s open-source implementation of SYCL, to demonstrate methods to achieve performant, portable code across five platforms available on the Intel Devcloud.
There are four modules within the series.
Module 1 Introduction to Performance, Portability, and Productivity for HPC This session will introduce Data Parallel C++ and the importance of performance, portability, and productivity for HPC development. There will be a Jupyter Lab environment for training, which will allow hands-on compilation and execution of simple DPC++ code samples.
Module 2 Optimization Best Practices Using SYCL/DPC++ This session will present a DPC++ code walk-through of simple matrix multiplication example, and look at how to optimize the code for performance using various features of DPC++ and to compare the execution times of various optimizations.
Module 3 Using Analysis Tools for Portable Offload to CPU or GPU This session will examine how to optimize the code for performance portability to get best performance regardless of whether the code is offloaded to GPU or CPU, and look at Intel VTune and Intel Advisor Roofline analysis for various optimizations.
Module 4 Mini Hackathon Attendees are encouraged to bring their own code; staff will be on hand to assist.