from the what-whirl-they-come-up-with-next? dept.
3-D Magnetic Interactions Could Lead to New Forms of Computing
A new form of magnetic interaction which pushes a formerly two-dimensional phenomenon into the third dimension could open up a host of exciting new possibilities for data storage and advanced computing, scientists say.
In a new paper published today in the journal Nature Materials, a team led by physicists from the University of Glasgow describe how they have been found a new way to successfully pass information from a series of tiny magnets arrayed on an ultrathin film across to magnets on a second film below.
Their breakthrough adds both a literal and metaphorical extra dimension to 'spintronics', the field of science dedicated to data storage, retrieval and processing, which has already had a major impact on the tech industry.
[...] At the nanoscale—where magnetic materials can be just a few billionths of a metre in size—magnets interact with each other in strange new ways, including the possibility of attracting and repelling each other at 90-degree angles instead of straight-on.
[...] The benefits of these 'spintronic' systems—low power consumption, high storage capacity and greater robustness—have made invaluable additions to technology such as magnetic hard disk drives, and won the discoverers of spintronics a Nobel prize in 2007.
However, the functionality of magnetic systems used today in computers remains confined to one plane, limiting their capacity. Now, the University of Glasgow-led team—along with partners from the Universities of Cambridge and Hamburg, the Technical University of Eindhoven and the Aalto University School of Science—have developed a new way to communicate information from one layer to another, adding new potential for storage and computation.
Dr. Amalio Fernandez-Pacheco, an EPSRC Early Career Fellow in the University's School of Physics and Astronomy, is the lead author on the paper. He said: "The discovery of this new type of interaction between neighbour layers gives us a rich and exciting way to explore and exploit unprecedented 3-D magnetic states in multi-layered nanoscale magnets.
"It's a bit like being given an extra note in a musical scale to play with—it opens up a whole new world of possibilities, not just for conventional information processing and storage, but potentially for new forms of computing we haven't even thought of yet."
[...] The team's paper, titled 'Symmetry-Breaking Interlayer Dzyaloshinskii-Moriya Interactions in Synthetic Antiferromagnets', is published in Nature Materials.
A paper is available on arXiv:1810.01801v2.
(Score: 2) by takyon on Wednesday June 05 2019, @10:50AM
Sounds like a memristor [wikipedia.org] equivalent.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 4, Interesting) by VLM on Wednesday June 05 2019, @12:27PM (2 children)
I can think of two historical generations of magnetic computing and they had some issues that might be interesting to think about WRT this new generation.
(Note, massive hand waving to simplify and make it brief, I know the following are not perfectly technically correct but are mostly not entirely wrong...)
The magnetic amps worked by finding a highly non-linear core, making a small signal transformer out of it, adding a third winding, and then you can turn the transformer on and off again by energizing the third winding perhaps with a very low current (wound many turns to make a large flux). So your transformer core overloads at flux level X, with the low current control winding off its a great transformer for signals of level 10% of X, but energize the control winding and that dude shuts off because (insert messy ferromagnetism stuff here). Historical problems were material variation, huge inductances are not fast electrically, some materials (not all) were temp sensitive.
The bubble memory idea was people are pretty chill about reading and writing patterns to magnetic tape, but it turns out you can induce (oh the pun) patterns to wander along a material kinda like mag tape "print thru" (which only gen-x and boomers will have experienced, LOL) so make a loop and store data by letting the patterns move along on their own. Clocking was an issue, temp, note pretty much vibration and radiation proof for aerospace. Also storage density was good by 70s standards, miserable by 80s standards. Unfortunately it took until the 80s to complete R+D and start mass production, whoopsie.
I'd predict future generations of magnetic computing gadgets to have similar problems, material quality, temp stability, inductance too high for fast I/O, low density. I'm sure it'll find a weird niche in ICBM IMUs or whatever, but probably not on our desks.
(Score: 2) by takyon on Wednesday June 05 2019, @01:42PM (1 child)
As traditional CMOS continues to race towards a brick wall, we will see more funding and interest in these alternative forms of computing.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by VLM on Wednesday June 05 2019, @01:57PM
True, but the argument against is not all forms are of the family of physics historically scaled, so its not looking all that good for the latest try. Hopefully it'll work this time.
By analogy of the magnetic computing, personally I like the idea of micromachine implementations of relay based computers, that would be so cool to have little nanometer scale telegraph relays. Probably wouldn't work, but it would be fun to think about.
(Score: 0) by Anonymous Coward on Wednesday June 05 2019, @04:32PM
row hammer++