Using Smoke and Mirrors to Compile a Functional Programming Language to Efficient GPU Code
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only! Grab it Explore the challenges and innovative solutions in compiling functional programming languages for efficient GPU execution in this 44-minute conference talk from Lambda Days 2022. Delve into the world of parallel programming as Troels Henriksen, Assistant Professor at the University of Copenhagen, unveils the secrets behind designing a functional language that can be transformed into hardware-friendly code. Learn about the Futhark language, which combines the elegance of functional programming with strategic limitations to generate GPU code competitive with hand-written alternatives. Discover crucial transformations, compiler techniques, and memory layout strategies that enable programmers to write clean, parallel, purely functional code while achieving high performance on GPUs.