[NEWS] - Infineon produces MRAM prototype

[CougTek]

This could be the beginning of the end for mechanical hard drives.

Unlike conventional computer chips, MRAM chips use magnetization to store information, instead of electrical charges. Up until now, computers had to continuously supply the main memory chips with power in order to store data. The information on MRAM memories, however, remains intact even after the system has been powered down. The working concept is similar to the way that data is stored on a computer’s hard disk drive. In addition, they are very fast, which enables exceptionally rapid access to data. For comparison: Infineon’s 16-Mbit MRAM chip is approximately 1,000 times faster than the non-volatile flash memory used in USB sticks, handheld computers and digital cameras. Furthermore, information can be stored in an MRAM memory cell up to one million times more often than is possible with flash memory.

The high storage capacity of the 16Mbit Chip (16 million single cells) was enabled using a small 1.4 square microns cell. By comparison 5,000 storage cells would fit on a cross section of a human hair.

News source

Submitted by JTR1962

 

[P5-133XL]

We're back to Core memory: What is old becomes new, once again.

P.S. Yes, I understand that it really isn't core at all but a much more sophisticated magnetic ram. But then I wouldn't have had anything pertinent to say.

 

[CougTek]

By the time the manufacturing process becomes mature and the size of the cells allow MRAM to compete with hard drives, I think MXO/WDC/STX and friends still have plenty of time before them.

 

[jtr1962]

I talked about this a lot on SR. While I think mechanical hard drives have roughly ten years of life left in them, there are a whole host of reasons why they're on the way out. Speed (or rather lack of it), noise, vibration, lack of shock resistance, power consumption are a few. The ability to scale beyond a few TB (my opinion) in capacity is another.

I think these will first replace flash RAM once they approach its current cost and capacity points. Within probably five years you'll have SSDs of maybe 10 GB suitable for boot drives. I'd say within ten years MRAM or something similar will exceed mechanical drives in capacity, and cost less per GB.

The HDD manufacturers should hedge their bets and get in on solid-state storage R&D now. Whether MRAM or some new technology, ten years or twenty, it's only a matter of time before mechanical disks go the way of the dodo, along with internal combustion engines and incandescent lights

 

[Santilli]

Oh, I really hope so...

s

 

[Will Rickards]

I think they'll have problems shrinking it.
Hard drives will still be around for a long time for larger storage and much higher rewrite ability. I think it may allow an almost instantaneous boot drive. This will actually be the coolest thing as computers will approach the same instant on as the kitchen appliance.

 

[Buck]

What's wrong with the internal combustion engine?

Anyway, I welcome the idea of MRAM as a replacement for the mechanical storage devices we use now. This would also replace optical drives as you could have very large capacity memory cards that fit in your shirt pocket.

 

[CougTek]

What's wrong with the internal combustion engine?

If you're talking about your own internal combustion engine, I would say that smell is the main problem.

 

[Buck]

You can smell that all the way over in that Belgian province of yours?

 

[CougTek]

New development

I really hope this comes to drives on shelves ASAP. RAM-like performance and good capacities.

 

[mubs]

I hope so too. Too many of these fancy things stay in the lab.

Took me a few moments to eralize this threaded was started in 2004; I thought it was a new one! Wonder what happened to Infineon's work.

 

[LunarMist]

New development

I really hope this comes to drives on shelves ASAP. RAM-like performance and good capacities.

ASAP may be many years yet.

 

[Bozo]

Took me a few moments to eralize this threaded was started in 2004; I thought it was a new one! Wonder what happened to Infineon's work.

Same here!

 

[Mercutio]

I'd guess that the project got killed because the projected ROI wasn't high enough to continue to make the R&D investment.

 

[Tannin]

I thought it was current right up until JTR started talking about developing 10GB SSD boot drives .... then I started to smell a rat.

 

[jtr1962]

I thought it was current right up until JTR started talking about developing 10GB SSD boot drives .... then I started to smell a rat.

Most of my predictions from 2004 did come true. We did have 10+ GB SSD drives in 2009, and so far we've been unable to scale mechanical drives past a few TB.

It actually looks likely that memristors, not MRAM, will be the next big thing. MRAM may yet replace DRAM, although I doubt it will scale enough to be useful for bulk storage.

I'm predicting petabyte memristor drives by 2020 (and the demise of spinning disks a few years before that). Let's see if these predicitions hold out in eight years.

 

[Tannin]

^ You are doing great so far. Any chance you can turn you attention to the short-term future of Randwick? I'd particularly like to know if Laughing Boy will be any good over the 10,000 metres in the fifth, and whether you reckon I should back him on the nose or each way.

 

[jtr1962]

Nope, no luck for me predicting sporting events. Even things I follow closely like the Tour de France I get wrong more times than I get right. Horses are even harder to predict than cyclists given that a bad day by either the horse or jockey can totally spoil the outcome.