Blog post: Optimizing Cubed

I’ve written a blog post about the work I’ve been doing this year in collaboration with @rabernat and @TomNicholas to improve the performance of Cubed, a library for serverless computations on large multidimensional arrays: Optimizing Cubed.

TL;DR: We implemented a number of optimizations in Cubed to give a 4.8x performance improvement on the “Quadratic Means” problem when running on Lithops with AWS Lambda, with a 1.5 TB workload completing in around 100 seconds.