Energy-free superfast computing invented by scientists using light pulses

Energy-free superfast computing invented by scientists using light pulses
Using ultrashort pulses of light enables extremely economical switching of a magnet from one stable orientation (red arrow) to another (white arrow). This concept enables ultrafast information storage with unprecedented energy efficiency. Credit: © Brad Baxley (parttowhole.com)

Superfast data processing using light pulses instead of electricity has been created by scientists.

The invention uses magnets to record computer data which consume virtually , solving the dilemma of how to create faster data processing speeds without the accompanying high costs.

Today's data centre servers consume between 2 to 5% of global electricity consumption, producing heat which in turn requires more power to cool the servers.

The problem is so acute that Microsoft has even submerged hundreds of its data centre services in the ocean in an effort to keep them cool and cut costs.

Most data are encoded as binary information (0 or 1 respectively) through the orientation of tiny magnets, called spins, in magnetic hard-drives. The magnetic read/write head is used to set or retrieve information using electrical currents which dissipate huge amounts of energy.

Now an international team publishing in Nature has solved the problem by replacing electricity with extremely short pulses of light—the duration of one trillionth of a second—concentrated by special antennas on top of a magnet.

This new method is superfast but so energy efficient that the temperature of the magnet does not increase at all.

The team includes Dr. Rostislav Mikhaylovskiy, formerly at Radboud University and now Lancaster University, Stefan Schlauderer, Dr. Christoph Lange and Professor Rupert Huber from Regensburg University, Professor Alexey Kimel from Radboud University and Professor Anatoly Zvezdin from the Russian Academy of Sciences.

They demonstrated this new method by pulsing a magnet with ultrashort light bursts (the duration of a millionth of a millionth of a second) at frequencies in the far infrared, the so called terahertz spectral range.

However, even the strongest existing sources of the terahertz light did not provide strong enough pulses to switch the orientation of a magnet to date.

The breakthrough was achieved by utilizing the efficient interaction mechanism of coupling between spins and terahertz electric field, which was discovered by the same team.

The scientists then developed and fabricated a very small antenna on top of the magnet to concentrate and thereby enhance the electric field of light. This strongest local electric field was sufficient to navigate the magnetization of the magnet to its new orientation in just one trillionth of a second.

The temperature of the magnet did not increase at all as this process requires energy of only one quantum of the terahertz light—a photon—per spin.

Dr. Mikhaylovskiy said: "The record-low energy loss makes this approach scalable.

Future storage devices would also exploit the excellent spatial definition of antenna structures enabling practical magnetic memories with simultaneously maximal energy efficiency and speed."

He plans to carry out further research using the new ultrafast laser at Lancaster University together with accelerators at the Cockroft Institute which are able to generate intense pulses of to allow switching magnets and to determine the practical and fundamental speed and energy limits of magnetic recording.


Explore further

Low energy electric field found suitable for quick magnetic recording

More information: Temporal and spectral fingerprints of ultrafast all-coherent spin switching, Nature (2019). DOI: 10.1038/s41586-019-1174-7 , https://www.nature.com/articles/s41586-019-1174-7
Journal information: Nature

Citation: Energy-free superfast computing invented by scientists using light pulses (2019, May 15) retrieved 19 May 2019 from https://phys.org/news/2019-05-energy-free-superfast-scientists-pulses.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.
2114 shares

Feedback to editors

User comments

May 15, 2019
This is a great new hard drive, but not a computer.

May 15, 2019
energy-free
Yeah. Typical picosecond pulsed laser, 40 mHz, 40 watts.
_______
Samsung 850 Pro
Available Capacities: to 1 TB
550 MB per second read (256 GB).

That's 4400 megabits/s ~5 watts.

May 15, 2019
This is a great new hard drive, but not a computer.


Depends on how you use it. Magnetic spins can interact, so you could implement logic gates where simply writing the data causes simple computation to take place. This would allow massive parallel processing of simple data as bitmaps.

May 15, 2019
simply writing the data
There's no 'simply writing' with a real-world picosecond-pulse teraHz laser writing at 40 megabits/s. It doesn't make any difference of the read/write head itself is an energy-miser. The whole thing has to be factored in.

May 15, 2019
. . inadvertent copy deleted.

May 15, 2019
The real question here is can we build nano-emitters in the THz range. Maybe quantum dots?

May 16, 2019
The title is frankly atrocious. Computing (or *any* kind of work) cannot be done "energy free". That would violate the laws of physics, since it would be the computing equivalent of a perpetual motion machine. This technique is very power efficient but it is nowhere close to being "energy free".

Neuromorphic computing chips like IBM's TrueNorth are also very power efficient (they operate with ~1/10,000 the power of a conventional CPUs, i.e. with a few tens of milliwatts). Thus they also do not heat up and require no cooling at all, but there is nothing "energy free" about them either. I am unsure if the title is a product of ignorance or intended as clickbait. I would guess it's the latter.

May 16, 2019
Considering that this would not only reduce hard drive usage but obviate dynamic RAM rewrites, it has the potential to greatly reduce power consumption. With femtosecond reads and writes, it also obviates the need for fast memory needed near the core, to a great extent. The only delay to memory would be foot-per-nanosecond transmission delay; with fast enough processors this would be sufficient to site memory close to the core, but the advantages would be greatly reduced.

On the other hand, this isn't going to be implemented soon. As usual, I would estimate at least a decade to develop this technology into working chips in a commercial computer.

May 16, 2019
They aren't talking about information processing here but rather information storage and retrieval i.e. they are talking about computer memory. The "Energy-free superfast computing" part of the title of this article is not only highly misleading (because it makes it sound like they are talking about information PROCESSING) but also the "Energy-free" part is just nonsense (That should have been "Energy efficient" because energy still used). Its the editor's fault that it was made so misleading and this shouldn't be taken as a reflection of the actual research which is excellent.

May 16, 2019
Sadly, its all too typical of journalist and editors to mar excellent scientific research with highly misleading wording or even lies in the articles they make public about it. This is highly damaging to science and often gives the false impression that the excellent scientific research is actually bad and/or 'nonsense' research. Sadly, I have seen this in many science articles.

May 16, 2019
@humy:
FYI, editors, proof readers and fact checkers are now three extinct species.
They died off about 30 years ago, as demonstrated here and any other form of media.

May 16, 2019
I think the issue with the terrible headline (and others like it) lies with someone at phys.org...because the article can be found verbatim on many sites - only the headline differs each time.

E.g. on "sciencedaily" the headline reads
Virutally energy-free superfast computing invented by scientists using light pulses


On "computing.co" the headline is
Superfast data processing speeds with near-zero energy consumption demonstrated using magnets and light pulses

May 16, 2019
Is there a better alternative to phys.org? In my experience phy.org has the fewest of these asinine headlines and generally has better content. However the second I catch wind of a site that has less inflation, ignorance, or out right lying, in both the headlines and the content I don't plan on ever visiting phys.org again.

I'm also confused about who they think their target audience is verse who it actually is. I'm probably one of the least educated people in the sciences that frequents this site (that comment anyway) and I often find myself groaning, every time i reread some broken analogy or metaphor used to describe a relativity basic concept that was already explained to me quite well over a decade ago in my southern state public school.

I don't know how many more times my brain can handle something like "the universe expands like a balloon being filled with aiiiirrrr"

May 16, 2019
Is there a better alternative to phys.org?


Phys.org rarely does any editing or authoring. They just replicate the articles verbatim. This is a news aggregating site. The problem is at the source. For example, in this case the bottom of the article reads:

Provided by Lancaster University


If you want real reporting, you should subscribe to a magazine or journal. The issue there is that people take so long to research a topic, and readers aren't interested in that sort of depth, so the magazines too are full of popularized fluff. "Science" with a capital S, just like the TV version made for the lowest common denominator.

The journals are better reading, but it's hard stuff for the layman to understand because you get the uncensored and unaltered version that is usually hard to understand unless you too are intimately familiar with the field.

May 16, 2019
or out right lying, in both the headlines and the content I don't plan on ever visiting phys.org again.

None of the articles on phys.org are authored by phys.org.
This is an aggregation site. All they do is reformat from press releases taken from elsewhere and add a headline.

So if you think that 'lying' is an issue on phys.org you haven't really grasped what this site is in the 6 months you're here.

Bravo

*slow clap*

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more