Resistance is futile? Memristor RAM now cheap as chips

Cheaper memristors could result from an accidental discovery at University College London. The HP-popularised memristor device is a form of ReRAM – resistive RAM – and is fairly expensive to make. Metal oxide-based ReRAM technology promises to combine minimum memory speed with NAND non-volatility and be able to provide higher …

[Anonymous Coward]

About time. Seems to be very little point in spending a lot of money on an SSD until they make them from memristors. Although it may be 5-10 years.

[Michael Strorm]

"Seems to be very little point in spending a lot of money on an SSD until they make them from memristors. Although it may be 5-10 years."

Are you serious? Storage technology (HDD or SSD) is generally rendered obsolete or out-of-date faster than that anyway, but it doesn't stop people buying it.

By the point "5-10 years" in the future these devices might- or might not- be available, the SSD you "spent a lot of money on" today will *already* likely have been rendered unimpressive by much faster and cheaper developments of regular storage- and they'll probably have been replaced and discarded at least once, if not twice anyway.

If this technology pans out and comes to fruition, then great! I wouldn't put off my own SSD purchase today though.

Feel free to put off *your* purchase though, and be prepared to wait further if- as often happens- a promising technology hits a roadblock. I do hope you didn't put off buying a floppy drive in the early-80s because bubble memory was on the horizon. :-)

[Luke McCarthy]

Endurance

Endurance: 3000? That's the number of rewrite cycles, right? I guess that still needs work...

[Nigel 11]

Re: Endurance

It's not a whole lot worse than TLC flash!

With decent controller architecture and a lot of spare bits, you could today make a decent memory stick or SSD out of this. OK, you'd not want to use it as cache or index storage for a busy database.

If it has small pages instead of huge ones like flash, and if that 3000 is worst-case not average life, it might actually be better than TLC already.

[Chris 155]

Re: Endurance

People always look at this and thing "Oh, that's no good", but what that means is that with a half way decent controller, you can write the full volume of this disk 30000 times. So if you're writing the full volume of your disk every day(which would be high utilization it'll last approximately 8.2 years or about twice as long as your HDD.

There are really just very few cases where this level of rewrite cycles is even remotely an issue.

[Dazed and Confused]

This is going to be fun

This is going to be fun

Since HP own the fundamental patents on this technology I can just see the EULA banning the use of these device to support any Oracle product.

If you scratch my eyes out, I'll scratch your's out.... Bitch

[Destroy All Monsters]

Re: This is going to be fun

I don't think HP owns the patent to a new production process of a passive device described in the 70s.

[h4rm0ny]

Re: This is going to be fun

Do you have a reference for HP's patent? I would be surprised if it covered this, tbh.

[Ian Michael Gumby]

Re: This is going to be fun

It's a new take on the production technique which means HP doesnt hold the patent.

Of course there is a lot more work to be done...

[Anonymous Coward]

There is always a catch...

...with new tech so when they finally get all the bugs worked out, then and only then will SSD and Memristors be viable for those who need secure data.

[wiggers]

"huge range of potential applications"

Boom boom! Basil Brush would have been proud...

[Bronek Kozicki]

holy sh* ...

10ns latency nonvolatile memory storage device, at price point of SSD ?! I want it yesterday!

[Ron 6]

"minimum memory speed"

Isn't zero iops the minimum memory speed?

[Graham Wilson]

At last.

Here's my order.

[Steve Briggs]

A ways to go? More like several continents away

The dimensions of the device make it a total non-starter 12500 x 12500 nm:??? WUWT?

I'd like to be hopeful, but that's going to require a lot of shrinkage to get to competitive.

[Francis Boyle]

Re: A ways to go? More like several continents away

WUWT? It's still in the laboratory. There's no reason to make it small until you need to make it small. That would just be making life difficult for yourself.

[4ecks]

Success from failure

In the best traditions of science and invention.

Failure to make one thing leads to understanding and discovery of another.

[jon 72]

accidental discovery my arse

Mines the one with the hardback copy of 'Tommorrows World' published by the BBC that describes a remarkeably similiar technology.

[The Indomitable Gall]

Re: accidental discovery my arse

Erm... that doesn't preclude accidental discovery. Temporary adhesives existed long before the Post-It note, but the specific adhesive in Post-Its was accidentally discovered while 3M were trying to invent a rival to Loctite's SuperGlue.

The point is that the memrister existed, but was expensive to make. This is an apparently cheaper way to make one, and it was discovered by accident.

[Bush_rat]

All well and good.....

But can it run crysis?

[TeeCee]

Re: All well and good.....

No, but you'll be able to find out whether your machine can run it that much more quickly.

[James Pickett]

"minimum memory speed"

I think he means 'maximum'...

[John 62]

dropped the wafer?

Is it just me or does it look like that wafer used be be one be circle?

[Juan Inamillion]

Why

Is there an utterly pointless picture of the alleged device? Like when pictures are put up of some bloke in a clean-room suit holding a disc of wafers... What is the unaided eye supposed to see? You could stick any old stock picture up and no one would know any difference... <grumble grumble>

/Sir Herbert Gussington-Smythe