Get ready for a crazy fun & informative talk! Boost your C++ code's performance by utilizing the full power of your computer’s built-in cache. In this talk, we'll delve into the basics of caches, spatial and temporal cache locality, and practical techniques for optimizing cache access. We'll also explore cache prefetching, branch prediction, and false sharing, providing examples modelled on real-world production code and performance benchmarks to drive home each point. By the end of this talk, you'll be equipped with the knowledge and tools to better optimize your code for cross-platform performance gains.
Michelle Fae D’Souza is a Software Engineer at Bloomberg, where she develops C++ code for the company’s order-trade entry and modification (OTE API) solution, which streamlines trading activity for many firms in the financial world. She is an active member of Bloomberg's C++ Guild and is a Technical Rep at Bloomberg. Michelle holds a bachelor’s degree in computer science from the University of California, Berkeley, where she was a Davis Scholar and President of the Computer Science Honor's Society.