Special Purpose Computers for Statistical Physics: achievements and lessons
Lev N. Shchur

TL;DR
This paper reviews the development and achievements of specialized computers for statistical physics in the late 20th century, highlighting their innovative design, superior performance, and scientific impact.
Contribution
It introduces a novel approach to specialized computer design optimized for integer operations, leading to unprecedented computational speed and accuracy in physics simulations.
Findings
Computers achieved three orders of magnitude faster calculations than supercomputers.
They enabled new scientific results with higher accuracy than previous methods.
The paper discusses lessons learned from long-term scientific computing projects.
Abstract
In the late 80s and 90s, theoretical physicists of the Landau Institute for Theoretical Physics designed and developed several specialized computers for challenging computational problems in the physics of phase transitions. These computers did not have a central processing unit. They optimize algorithms to handle elementary operations on integers -- read, write, compare, and count. The approach allowed them to achieve recording run times. Computers performed calculations three orders of magnitude faster than similar calculations on the world's best supercomputers. The approach made it possible to obtain fundamentally new results, some of which have not yet been surpassed in the accuracy of calculations. The report will present the main ideas for the development of specialized computers and the scientific results obtained with their help. The lessons of planning and execution of…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
