Gagundathar, thought you’d be interested in this…
As a computer scientist/engineer, how do you see this impacting your field of expertise?
None, really. The article mentions quantum computing - which would change things - but so far as I could tell this was simply a reduction in transistor size, not an implementation of quantum computing.
I think people are already, for better or worse, presuming that Moore’s law will continue. This just tells us that their presumption is likely fairly safe, but doesn’t change what they’re assuming.
If the feature size is only 1 atom (and it works reliably), does that mean that chip designers can pack up and go home? (Well, I guess we would like to figure out how to make 3D chips if they haven’t already and how to make them zero-cost and zero-power, but other than that…)
In other words, does a one atom transistor mean the end of Moore’s Law?
Thanks,
Rob
The time between researchers demonstrating a principle and products being engineered is pretty long. This sounds suspiciously close to a Diamond Age kind of technology.
Nah. Subnuclear quark-computing comes next. It’s gotta.
The article is very unclear as to if this is a step to quantum computing or the ultimate in normal circuit design. It is not at all clear to me how this thing switches.
20 years from now is not very long, says someone who learned transistor theory 40 years ago. But there are a lot of issues beyond just manufacturing.
First, this only seems to work at very low temperatures, so you wouldn’t want to carry your iPhone too close to anything important. I suspect this is done to keep the atoms quiet.
The big thing is that random cosmic ray hits are causing havoc already. Memories are designed to recover from these, but we are getting more and more worried about the impact on logic. It is common for people to make a pilgrimage to Los Alamos to zap a bunch of test chips and see how it holds up to radiation. Think if what happens when you are dealing with single atoms and equally small signal paths.
You are probably going to have one atom transistors but an actual transistor will be made of several atoms, distributed, and some sort of a voter. You can’t make them close together because one hit might disrupt several in the same way so the voter will think that the bad value is good. Or you can put the entire very cold and small chunk of logic in a lead casing.
As for me, in 20 years I’ll be long retired, so I can read papers about this stuff and enjoy not having to deal with it.
What are “papers,” grandpa?
Who said that papers come out on paper any more?
I am pretty sure papers don’t exist. They say they do, but there’s this huge paywall in front of everything.
In the future computers will be the size of a dust mote, but will require a containment vessel.
Rob
Moved to IMHO from Great Debates.