Loop Rerolling for Hardware Decompilation

Explore a groundbreaking 19-minute video presentation from PLDI 2023 that introduces the concept of hardware decompilation and focuses on a novel technique called hardware loop rerolling. Delve into the research conducted by Zachary D. Sisco, Jonathan Balkind, Timothy Sherwood, and Ben Hardekopf from the University of California at Santa Barbara, which aims to recover higher-level programming abstractions from low-level netlists. Learn how this technique leverages clone detection and program synthesis to identify repeated logic in netlists and reroll them into syntactic loops in recovered HDL code. Discover the evaluation results across various hardware design benchmarks, demonstrating successful loop identification and rerolling in 52 out of 53 netlists. Gain insights into the practical benefits of hardware decompilation, including HDL transpilation, faster simulation times, and artifact compaction. Access supplementary materials and explore the potential impact of this research on the field of hardware design and analysis.