Heterogeneous application development has introduced new challenges for experienced and novice developers alike. In particular, the need to understand how your programming language maps algorithms to each device.
This webinar focuses on how to gain that understanding for cross-architecture applications written in Data Parallel C++ (DPC+). Specifically, it showcases how Flow Graph Analyzer, a feature of Intel® Advisor, provides insights into efficiencies of DPC++ data parallel constructs that help you improve algorithm performance.
Join Intel software specialists Anoop Madhusoodhanan Prabha and Kevin O’Leary to learn about:
- The asynchronous task graph of your DPC++ algorithms and its similarities with other commonly used heterogeneous programming models
- How to identify performance penalties using diagnostics reported on simple language use
- How to determine scheduling inefficiencies—which kernels can be executed independently or if the kernels are being scheduled as intended
- Flow Graph Analyzer reports and recommendations that can help improve the performance of data parallel algorithms
Includes demonstrations and code walkthroughs.
Get the software
- Download Intel® Advisor as part of the Intel® oneAPI Base Toolkit, a core set tools and libraries for developing high-performance, data-centric applications across diverse architectures.
- Already have the Base Kit … or want Intel Advisor as a standalone product? Get it here.
- Sign up for an Intel® DevCloud for oneAPI account—a free development sandbox with access to the latest Intel® hardware and oneAPI software.
- Explore oneAPI, including developer opportunities and benefits
- Subscribe to the POD—Code Together is an interview series that explores the challenges at the forefront of cross-architecture development. Each bi-weekly episode features industry VIPs who are blazing new trails through today’s data-centric world. Available wherever you get your podcasts.