Kx financial engineer, Connor Gervin, gave a talk at a Kx Community NYC Meetup in 2015 that is worth revisiting. Connor described kdb+’s built-in multithreading and multiprocessing capabilities, which are an essential part of every serious kdb+ programmer’s’ toolkit. With these features, programmers can make the best use of multicore hardware when solving increasingly complex problems over ever-expanding datasets.
In a guided overview with real code examples, Connor outlines how the multicore and multithreading aspects of parallel execution operations; multithreading input queues; multithreaded access to the HDB; segmentation and MapReduce all work together to remove memory limits and single-thread computing constraints.
Connor demonstrates a kdb+ solution to the complex problems posed by massive datasets that shows programmers how to significantly improve vector operations, disk performance and optimize queries.
Watch Connor’s presentation on the Kx Systems’ YouTube channel (here) and while you’re there, subscribe so you never miss new videos about Kx and kdb+.