Optimizing Dynamic-Graph Data Structures on Multicores with the Locality-First Strategy

Explore the challenges and strategies for optimizing dynamic-graph data structures on multicore systems in this 48-minute talk by Helen Xu at the Simons Institute. Delve into the complexities of developing efficient code for large-scale problems, ranging from gigabytes to terabytes, while leveraging multicore hardware features. Learn about the importance of balancing parallelism and locality in software systems to maximize performance. Examine the specific challenges posed by real-world dynamic graphs, including their sparse and skewed structures, and how these characteristics impact locality and parallelism. Gain insights into the locality-first strategy for optimizing dynamic-graph data structures, and understand its potential benefits in addressing the unique demands of multicore environments.