Long Live Small Fan-in Majority Gates Their Reign Looks Like Coming!

Ibrahim, Walid; Beiu, Valeriu

doi:10.1109/asap.2007.4429993

Cited by 1 publication

(2 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An immediate enhancement would be to use Monte Carlo simulations at the gate-level instead of using an equation [33]. • Device-level estimates of vN-MUX show much more complex behaviors than those revealed by gate-level analyses.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

A fresh look at majority multiplexing when devices get into the picture

Beiu

Ibrahim

Lazarova-Molnar

2007

2007 7th IEEE Conference on Nanotechnology (IEEE NANO)

Self Cite

View full text Add to dashboard Cite

In this paper we present the first detailed analysis of von Neumann multiplexing (vN-MUX) using majority (MAJ) gates of small fan-ins ǻ (MAJ-ǻ) with respect to the probability of failure of the elementary (nano-)devices. Only gates with small fan-ins have been considered, as gates with large fan-ins do not seem practical (at least in the short term) in future technologies. The extensions from an exact counting algorithm (for gate defects and faults only) to device-level failures will allow us to estimate and characterize MAJ-ǻ vN-MUX with respect to device-level malfunctions. The reported results depart significantly from all known gate-level analyses-either theoretical or based on simulations. These should be quite important as providing a detailed picture of the behavior of MAJ-ǻ vN-MUX when considering the (un)reliability of the elementary (nano-)devices (as opposed to gate-level only analyses). The main conclusion is that small fan-in gates (and redundancy schemes relying on such gates) are quite promising-in spite of all previous results at gate-level showing the contrary.

show abstract

Section: Discussionmentioning

confidence: 99%

“…We expect that more precise estimates than the ones presented in this paper will be possible when using Monte Carlo simulations in combination with higher level reliability algorithms. These are clearly needed and have started to be investigated [33].…”

Section: Discussionmentioning

confidence: 99%