Slashdot Mirror
Scientists Demonstrate Ultra-Fast Magnetite Electrical Switch
adeelarshad82 writes "Researchers at the U.S. Department of Energy's SLAC National Accelerator Laboratory recently demonstrated electrical switching thousands of times faster than in transistors now in use thanks to a naturally magnetic mineral called magnetite (abstract). The experiment is considered a major step forward in understanding electrical structures at the atomic level and working with recently identified electrical 'building blocks' called trimerons. The breakthrough could lead to innovations in the tiny transistors that control the flow of electricity across silicon chips, enabling faster, more powerful computing devices."
No, but it is relevant for nerds.
As for whether or not new technologies ever pan out... perhaps you should compare whatever computer it is you're using now against the one you were using in 1980.
I don't care if it's 90,000 hectares. That lake was not my doing.
Yes, it's a dupe, but at least this one has a link to an article that explains something about it.
Could be. I can see possibilities for extremely high frequency oscillators and mixers used in optical data transmission.
You can't make a tiny transistor when they have to use a laser to turn on/off the device. I'll bet the laser is a lot bigger than the feature sizes of transistors these days. i.e. not going to have as complex circuits as we have today.
So how long to charge that laser, what type of delay you to laser fires and are you using a slower technology to switch the laser on/off?
You can improve the rise/fall time of the transistor, great, but if you have a lot of latency from the control signal to the laser to circuit switching, it doesn't improve on propagation delay. Factor all that in, is it still switching faster than the plain old transistor?
Does it count as prior art if the prior art is biogenic? If so... meh, this is old news!
Someone flopped a steamer in the gene pool.
Magnetite is already commonly used in magnetic storage such as hard disks.
Now that SSDs are replacing hard disks the magnetite suppliers are looking for new customers.
Want to bet they are going to fund the shit out of research into using it as transistors in household electronics?
I'm talking about the non-silicon THz frequency transistors that are promised every now and then, obviously the normal iterative approach is a valid approach for improvments, but it doens't lead to breakthrough paradigm shifts.
Is this actually relevant for end-user electronics? Or is it yet another of those wonderful promising potential fast-switching techs that are announced every few months(since 1980 or so) yet never pan out to anything practical.
It it's current form, no, at least not for desktops. It might be useful for supercomputers. Real supercomputers that is, not the supercomputers currently in vogue made of hundreds of pallet loads of commodity type PCs linked by networks. The requirement for cryocooling (-190 C.) pretty much rules it out otherwise.
Hopefully it will serve as a good starting point for further research that could lead to breakthroughs that allow it to work at higher temperatures.
much of left-wing thought is a kind of playing with fire by people who don't even know that fire is hot - George Orwell
But it was duped ultra-fast keeping in the spirit of the story.
this is normal for scientific R&D for possible future products, most things don't pan out. those very few things that get invested in don't pan out. most start up businesses don't pan out. One of my past jobs was manager of engineering group at profitable company, and even then most things done in R&D there don't pan out.
so don't complain, it's normal and always has been
It's made from Tight Magnets . . .
We are Dead Stars looking back Up at the Sky
scientists have reaffirmed that magnetite has absolutley ZERO PRACTICAL value in transistor applications. "No way in hell" was the unatributed quote.
The issue is that generally, something might be possible in the lab but impractical or tooexpensive to scale to commercial production levels. So, instead of a 100x performance increase, they might be able to use the information to give us a 10x performance increase, over the course of years of iterative development.
It is pitch black. You are likely to be eaten by a grue.
I think to myself, why don't you put it in a chip already so I can play games with real time ray-tracing at 4K resolution?! I wish I had technical knowledge to understand how far all these discoveries are from being implemented in a commercial CPU or a GPU. My ignorance frustrates me!!!
Is this actually relevant for end-user electronics? Or is it yet another of those wonderful promising potential fast-switching techs that are announced every few months(since 1980 or so) yet never pan out to anything practical.
You'd be surprised of how many of those things that already have found their way into your home but still pop up on slashdot because someone finds out some new production method to make them more viable in other application.
Take for example this article about GaAs semiconductors from 2001
You also have retarded comments like "Ah, Gallium Arsenide chips, thw chip of the future. Always have been, always will be, the chip of the future." from Blaede, a comment that reminds me of yours.
Yet today I'm pretty sure that the RF components in all your wireless devices are GaAs today. We had to replace our RF switches with GaAs a couple of years ago since the old technology was phased out.
Of course you don't see it because that would require opening it up and reading the datasheet for the components.
Technology and science aren't selling points so you will never see the technology written about if you only look at consumer pages and shiny packaging. There are other pages that highlights the news that you are interested in, like PCWorld and Macworld
Because raytracing does produce near-reality quality. It solves the shadow/lighting problem - you simply treat diffuse light sources as a very large number of point light sources. All those wonderfully-perfect CGI films and special effects are produced using raytracing... at an hour a frame.
The only problem with raytracing is the extreme processing requirements. It's paralleliseable to a point, until you start having issues with memory bandwidth.
Or that the commercial entity which ends up with the technology wants to get the most they can out of it commercially and release it as slow incremental improvements that put them just a little ahead of their competitors over and over again for years. If they go full on and put in the money to dish it out all at once they get a single big leap over their competitors for the time it takes their competitors to adopt said technology as well and then they have nothing else to release to answer back.
Yeah, stop wasting money on research. This stuff never pans out, so lets stop trying.
It gripped her hand gently. 'Regret is for humans,' it said.
s/idiots/people/
s/intellect/knowledge/
FTFY.
Take for example this article about GaAs semiconductors from 2001...
/. article. Man, how I miss the old /., when knowledgeable people posted technical stuff worth reading!
Thank you, AC, for posting this interesting old
Circle the wagons and fire inward. Entropy increases without bounds.