3 ms·
A very entertaining and informative essay. Could perhaps someone try to elaborate on the calculation of buffers needed per chip? Quote: To a physicist this ma
by wuschel 8y ago
A very entertaining and informative essay.
Could perhaps someone try to elaborate on the calculation of buffers needed per chip?
Quote:
To a physicist this may seem natural, but to a computer designer, treating a set of boolean circuits as a continuous, differentiable system is a bit strange. Feynman's router equations were in terms of variables representing continuous quantities such as "the average number of 1 bits in a message address." I was much more accustomed to seeing analysis in terms of inductive proof and case analysis than taking the derivative of "the number of 1's" with respect to time.
How did he do it, conceptionally, and in practice?
- aeternus 8y agoThe problem sounds like it can be solved with queueing theory. The (common) inductive proof method is covered by the wikipedia article: https://en.wikipedia.org/wiki/Queueing_theory#Example_analysis_of_an_M/M/1_queue https://en.wikipedia.org/wiki/Queueing_theory#Example_analys... However since the arrival of messages can be modeled as a probabilistic or continuous rate, differential equations can also be used. I'm not sure if this is how Feynman did it, but this paper shows how to model queueing using differential equations: http://pi.math.cornell.edu/~rand/randpdf/ENOC2017_queueing.pdf http://pi.math.cornell.edu/~rand/randpdf/ENOC2017_queueing.p...
- wuschel 8y agoThanks for looking into this, and posting these directions!