Designing a Slimmer Vector of C++ Variants

Explore an in-depth conference talk from C++Now 2024 where senior Bloomberg LP engineer Christopher Fretz presents innovative solutions for optimizing memory usage in heterogeneous containers. Learn how to address the memory inefficiencies of std::vector when dealing with mixed types of different sizes, particularly when larger types are infrequently used. Discover the design principles behind a bit-packed alternative data structure that maintains the familiar API while significantly reducing memory overhead. Gain insights into advanced C++ container design, memory optimization techniques, and practical solutions for handling variants in performance-critical applications. The presentation includes detailed explanations, implementation strategies, and real-world applications that demonstrate how to achieve substantial memory improvements in heterogeneous container implementations.