Faster APL with Lazy Extensions

Explore a 36-minute conference talk from ARRAY 2023 that delves into optimizing APL performance through lazy evaluation techniques. Learn how the April compiler, which translates a subset of APL to Common Lisp, implements deferred evaluation to create more efficient and elegant APL code. Discover the concept of 'virtual arrays' and how they build a tree of objects representing arrays not yet computed. Understand the two-stage compilation process that generates high-performance array-building kernels, with the potential to emit assembly code for maximum speed. Gain insights into how this evaluation model leverages object-oriented programming in Common Lisp to achieve both performance and elegant specification. Presented by independent researcher Andrew Sengul, this talk offers valuable knowledge for those interested in APL, Lisp, just-in-time compilation, and advanced programming language implementation techniques.

[ARRAY'23] Faster APL with Lazy Extensions