IBM Breakthrough Exponentially Expands Data Storage

[DarkRyter]

Oh yeah, singularities and whatnot.

So bitchin.

 

[ReZerO]

I remember when i first got a hard drive that was more than one megabyte... that was epic at the time..

 

[Tharwen]

Well if we were to keep using discs the size of current Blu-Ray's and DVD's, how much data could we potentially cram onto one disc?

I'm fairly sure this technology only affects magnetic hard disks, not disks like that.

To be honest though, my knowledge is sketchy and unreliable.

 

[2fish]

I love the idea but 5-10 years? I suppose I will wait and be happy with my current set up but damn that sweet. Won't matter though I will still find a way to fill it with crap. I will always only have a few GB free. Someday I will clean out my media file... honest.

I almost wish that it was still necessary to fully download porn just so I could see how much I could store on it.

Start downloading free screensavers instead?

 

[Silas13013]

But what will our future be? HDD, or SSD?

What's our priority? Virtually zero load time or not having to uninstall a game, ever again?

DECISIONS! DECISIONS!

Both. Intel cache tech which allows you to have an SSD as your HDD cache. Best of both worlds, massive speed on the programs you use most often and the cheap storage of a HDD

 

[The_Echo]

And just yesterday I was impressed by a 16gig microsd card...

Seriously, this stuff is awesome.

They already go up to 64GB.

This news basically made my day. I'd absolutely love a 150TB hard drive.

 

[The_Blue_Rider]

And just yesterday I was impressed by a 16gig microsd card...

Seriously, this stuff is awesome.

They already go up to 64GB.

This news basically made my day. I'd absolutely love a 150TB hard drive.

Holy crap if I got a 64GB microSD my phone would have more memory than my computers Hard Drive D:

 

[LilithSlave]

I already use more than 10 TB of storage, given that the myriad of my several 2 TB HDDs it practically full by now.

Improvements are very welcome.

 

[Lawlhat]

I wouldn't really flaunt reaching the end of Moore's Law. It has economic implications.

 

[Paradukes]

Yes, this is cool, and all, but...

If this is the future, where's my god-damned flying car?

 

[pwnzerstick]

Sooooo... Super Blue Ray?... Purple Ray? idk

 

[FarleShadow]

I wouldn't really flaunt reaching the end of Moore's Law. It has economic implications.

You know, the one thing I never understood about Moore's law (Double performance every 2 years) was that if we do reach the end of Moore's law, in terms of an extremely cheap, extremely tiny chip doing extreme calculations, why nobody suggested that the uberchip be complimented by another uberchip sitting nearby?

 

[Xan Krieger]

Yes, this is cool, and all, but...

If this is the future, where's my god-damned flying car?

Flying cars are cool till people start flying drunk. Also it's one thing for someone to wreck on a road, it's entirely another for them to wreck in mid-air and come crashing down on your house.

OT: I can just imagine how cheap 1 or 2TB harddrives will become at that point. I already store my entire steam library on a 1.5TB harddrive, if I had a 10TB harddrive or larger I'd save up some money then try to fill it on steam.

 

[isometry]

I wouldn't really flaunt reaching the end of Moore's Law. It has economic implications.

You know, the one thing I never understood about Moore's law (Double performance every 2 years) was that if we do reach the end of Moore's law, in terms of an extremely cheap, extremely tiny chip doing extreme calculations, why nobody suggested that the uberchip be complimented by another uberchip sitting nearby?

Good question. You're right that Moore's law is often loosely stated as "double performance every 18 months", but a more precise version that deals with your question is "double the density of transistors on processors every 18 months."

So it says that now you can fit 2x transistors in the same space that you could fit only x transistors 18 months ago. Reaching the end of Moore's law would mean not being able to make the transistors any smaller (increasing the size of the whole chip as you suggest is an option in theory, but raises all kinds of issues with heat, power, internal and external couplings, manufacture, etc that would need to be solved).

 

[Idlemessiah]

5-10 years? Awesome. 5 years ago you'd be impressed with a 500MB memory stick. 10 years ago if you could fit enough tracks on an mp3 player for the ride to school you were a wizard.

 

[cookyy2k]

I wouldn't really flaunt reaching the end of Moore's Law. It has economic implications.

You know, the one thing I never understood about Moore's law (Double performance every 2 years) was that if we do reach the end of Moore's law, in terms of an extremely cheap, extremely tiny chip doing extreme calculations, why nobody suggested that the uberchip be complimented by another uberchip sitting nearby?

Good question. You're right that Moore's law is often loosely stated as "double performance every 18 months", but a more precise version that deals with your question is "double the density of transistors on processors every 18 months."

So it says that now you can fit 2x transistors in the same space that you could fit only x transistors 18 months ago. Reaching the end of Moore's law would mean not being able to make the transistors any smaller (increasing the size of the whole chip as you suggest is an option in theory, but raises all kinds of issues with heat, power, internal and external couplings, manufacture, etc that would need to be solved).

The problem with shrinking transistors any further is quantum tunneling. Any smaller and the tunnelling current basically fries them.

As for this new advance the theory is great and all but the cost will be huge, they may be available in 10 or so years but affordable? STMs are not cheap pieces of equipment to buy, maintain and use.

 

[Jodah]

I wouldn't really flaunt reaching the end of Moore's Law. It has economic implications.

You know, the one thing I never understood about Moore's law (Double performance every 2 years) was that if we do reach the end of Moore's law, in terms of an extremely cheap, extremely tiny chip doing extreme calculations, why nobody suggested that the uberchip be complimented by another uberchip sitting nearby?

Good question. You're right that Moore's law is often loosely stated as "double performance every 18 months", but a more precise version that deals with your question is "double the density of transistors on processors every 18 months."

So it says that now you can fit 2x transistors in the same space that you could fit only x transistors 18 months ago. Reaching the end of Moore's law would mean not being able to make the transistors any smaller (increasing the size of the whole chip as you suggest is an option in theory, but raises all kinds of issues with heat, power, internal and external couplings, manufacture, etc that would need to be solved).

This. Its not just doubling the performance that is crucial for Moore's Law. Hell we could quadruple it if that's all it meant. We would just have computers the size of station wagons in every home! Moore's Law is about doubling the performance with the same amount of space every 18 months. IE performance doubles but computers stay the same size.

 

[FarleShadow]

I wouldn't really flaunt reaching the end of Moore's Law. It has economic implications.

You know, the one thing I never understood about Moore's law (Double performance every 2 years) was that if we do reach the end of Moore's law, in terms of an extremely cheap, extremely tiny chip doing extreme calculations, why nobody suggested that the uberchip be complimented by another uberchip sitting nearby?

Good question. You're right that Moore's law is often loosely stated as "double performance every 18 months", but a more precise version that deals with your question is "double the density of transistors on processors every 18 months."

So it says that now you can fit 2x transistors in the same space that you could fit only x transistors 18 months ago. Reaching the end of Moore's law would mean not being able to make the transistors any smaller (increasing the size of the whole chip as you suggest is an option in theory, but raises all kinds of issues with heat, power, internal and external couplings, manufacture, etc that would need to be solved).

Well, my guess is the inter/external coupling and making stuff would be resolved as the chips reduced in complexity (Either by increasing stuff on the chip/motherboard so that only the 'meat' of the issue was actually processed, (big chip expensive, so motherboard has alot of the older, cheaper 'chips' to translate)).

I actually think heat would be the biggest issue (Power being a function of heat).

But I actually think, when we can stamp out a supercomputer onto a standard chip, that we'll use that technology to imbed the ability to process information into simple parts, so instead of a tiny processer needing huge heat-dissipation technology for our phones, the entire case will become the processor element. atleast maybe, I need to think about that abit more.

 

[SpAc3man]

They still had to cool everything down to a fraction above 0K to get it to work. They have said that to be practical from a manufacturing point of view they would have to use around 150 atoms per bit rather than 12.

Still an awesome leap forward in storage technology.

 

[isometry]

The problem with shrinking transistors any further is quantum tunneling. Any smaller and the tunnelling current basically fries them.

True, though we'll learn how to build transistors on the quantum scale long before we get scalable quantum computers.

It might involve a radical redesign of the semiconductor junctions we use now, for example some researchers have proposed building classical cellular automata with quantum dots:

http://nd.edu/~qcahome/

The bits are stored as confined electrons, and computations are done through the bits interacting electrostatically.

As for this new advance the theory is great and all but the cost will be huge, they may be available in 10 or so years but affordable? STMs are not cheap pieces of equipment to buy, maintain and use.

Agreed, the real innovation would be a fast method for reading and writing these kinds of ultra dense atomic memories. Parallel to the discussion about building bigger processors, achieving a huge storage capacity is as simple as keeping 100 1TB drives, the real breakthrough would be a 1 TB drive with the kind of access times we are used to seeing in RAM.