Algorithms vs Moore’s Law

Sherry and Thompson study 113 algorithm families since the 1940s — for medium and large problem sizes, 30%-43% improve faster than Moore’s Law.

Silicon power vs brain power. Image: LinkedIn.

Silicon power vs. brain power.

Yash Sherry and Neil Thompson from Massachusetts Institute of Technology studied 113 algorithm families since 1940s — that’s 57 textbooks and 1,100 research papers — and see how fast they have improved compared to hardware/Moore’s law: for medium to large problem sizes, 30% and 43% of them have improved faster than silicon.

Sherry, Yash, and Neil C. Thompson. 2021. “How Fast Do Algorithms Improve?” Proceedings of the IEEE, 1–10. https://ieeexplore.ieee.org/document/9540991

Algorithm Wiki: http://algorithm-wiki.org

Originally posted on LinkedIn.