(This webinar was on March 18, 2021. See below for a link to the video.)
It is frustrating to parallelize your code and not get much of a performance benefit from using multiple cores. Parallel processing always comes with an additional cost over the execution of the sequential version of the program and the OpenMP API is no exception to this. If the language constructs and features are not used wisely, this overhead negatively impacts the performance gain realized using multiple cores– and this may even degrade the performance.
Join us for a webinar about how to tune your OpenMP API code to get the best performance. This presentation will be delivered by Ruud van der Pas, Distinguished Engineer in the SPARC Processor Organization at Oracle. In his presentation, Ruud will give guidelines on how to get the best performance using the OpenMP API. Several real-world examples will be shown and discussed, and he will also be available to answer your questions.
Ruud has significant experience working with the OpenMP API and participates heavily in the OpenMP Language Committee. He speaks regularly about OpenMP performance: check out his presentation from SC’20.