Stanford Seminar - Generalized Reversible Computing and the Unconventional Computing Landscape

Channel:Stanford Online · Length:1:10:03 · Published:2017-10-19 · Views:15592

Watch on YouTube

Summary

A Stanford EE380 seminar by Michael P. Frank of Sandia National Laboratories on generalized reversible computing. It surveys unconventional computing approaches as transistor scaling nears its limits and energy efficiency stops improving.

Why HN recommends it

HN commenters cite it as an informative talk from a leading researcher in reversible computing, mainly in threads on reversible computers and energy efficiency.

his Stanford talk is extremely informative — HN, 2020

Computer architecture & CPUs · Semiconductors & chip manufacturing

Subtitles (excerpt)

welcome everybody this is an area that I have been researching for quite some time I basically started in the mid 90s I had a project the first project I was going to do for my PhD originally was on DNA computing and actually invented a scheme for universal computing using DNA chemistry and it turned out that because of the laws of thermochemistry the DNA computer we designed could either take forward steps or backward steps through the computation and so in order to make any forward progress it had to be reversible because otherwise there would be more ways to go backwards than to go forwards and so it would just tend to go backwards right so sort of motivated by this I started studying reversible computing in more depth and turned towards electronic implementations of reversible computing which were a lot more practical certainly in the near term so my talk today is going to basically be a brief survey of sort of the motivation in terms of that an end of semiconductor scaling which we just heard about and then we can sort of break down the sources of energy dissipation and computing and to energy that's dissipated at different levels in a computer design and what we find is that a lot of the proposals for unconventional models or styles of computing

On Hacker News

I think Michael P Frank is one of the more knowledgable people working on the subject, his Stanford talk is extremely informative ... [VIDEO] — jonbaer, 2020-07
For anyone interested in the topic I highly recommend checking out Dr. Michael P. Frank's work. This video is good one to start with [VIDEO] — gbrown_, 2021-02
According to [Michael Frank's video]( [VIDEO] ), there is an architectural overhead of ~500x that people can not get over (IIUC, this overhead only applies for gate based computing models). A reasonable gap between state-of-the-art gate based computing models and Landauer's limit should be 10^2-10^3. On the other side, reversible computing in general has 2-16 times computational overhead at the s… — GiggleLiu, 2021-03

Related videos