Linked by Thom Holwerda on Wed 8th Sep 2010 22:12 UTC, submitted by poundsmack
Hardware, Embedded Systems "The recent announcement that Rice University grad student Jun Yao has demonstrated a new memory device has created a stir. As well it should. The new device is non-volatile, offers fast sub-100 ns switching times, can be written 10,000 times and is fully compatible with current CMOS manufacturing processes. A 1,000 bit proof-of-concept chip has been built by a private company."
Order by: Score:
Comment by cb88
by cb88 on Thu 9th Sep 2010 05:19 UTC
cb88
Member since:
2009-04-23

1000 or 1024 bits?

That 10k write limit could be hit in 10 ms.... so yeah they need to fix that. Perhaps they could somehow get the chip to use oxygen from the air...to increase the number of times the oxide chain could be switched on and off.

Reply Score: 1

RE: Comment by cb88
by WereCatf on Thu 9th Sep 2010 05:46 UTC in reply to "Comment by cb88"
WereCatf Member since:
2006-02-15

1000 or 1024 bits?

Robin Harris has been in the business for quite a while so I assume he knows what he's writing about, and thus 1000 bits is probably correct. An odd number for us geeks but then again, there is no reason why everyone should follow the same rules we do, especially when creating a proof-of-concept of useless size.

That 10k write limit could be hit in 10 ms.

10k write limit may sound little but then again, these things won't be sold to consumers for use in desktop computers. In mobile devices you often do not write that much stuff on the device all the time. For example, a digital video camera usually writes data in very long, large amounts but not all that often. With a storage memory consisting of this SiOx memory and some wear-leveling mechanism you could still aim for perhaps 7000-8000 video sessions.

I am quite certain they will raise the limit with time but even as it is it's good enough for lots of uses as long as they're carefully selected and adapted for such. 100ns write speeds for permanent storage just is too good for many real-world applications to pass up.

Reply Score: 3

RE[2]: Comment by cb88
by lemur2 on Thu 9th Sep 2010 06:11 UTC in reply to "RE: Comment by cb88"
lemur2 Member since:
2007-02-17

With a storage memory consisting of this SiOx memory and some wear-leveling mechanism you could still aim for perhaps 7000-8000 video sessions.

I am quite certain they will raise the limit with time but even as it is it's good enough for lots of uses as long as they're carefully selected and adapted for such. 100ns write speeds for permanent storage just is too good for many real-world applications to pass up.


Exactly so.

Given features of permanent storage, 100ns write speeds, 5 nm wide (1/5th the feature size of the latest flash devices), 3D stack-ability, 2 terminal memory, in conjunction with a wear-levelling filesystem and high-speed interfaces:
http://en.akihabaranews.com/60205/storage/new-sd-4-0-memory-cards-s...
... all of this together means that the market for read-many-write-infrequently non-volatile storage is going to really hot up. We could even see the advent of SD cards and/or USB sticks out-performing terabyte hard drives.

Imagine a micro-video-jukebox: a few hundred HD movies stored on a high-speed SD card. Couple that with a low-power HD-capable player (with a HDMI output):
http://topnews.us/content/225446-orion-samsung-s-new-dual-core-arm-...
What would you call it? iMovies?

Edited 2010-09-09 06:16 UTC

Reply Score: 2

RE[2]: Comment by cb88
by cerbie on Fri 10th Sep 2010 06:25 UTC in reply to "RE: Comment by cb88"
cerbie Member since:
2006-01-02

Not sold to consumers? We have SSDs, and write cycles seem to have actually been going down, as they've gotten smaller.

Reply Score: 2

RE: Comment by cb88
by Ruahine on Thu 9th Sep 2010 05:51 UTC in reply to "Comment by cb88"
Ruahine Member since:
2005-07-07

A 10k write limit is not uncommon in non-volatile memory systems. A few years ago things were even worse. True, with this limit you're not going to be using it as RAM for your PC, but there are plenty of other potential uses for it. It could, for example, make a huge difference to the microcontroller market a few years down the line.

Edit:
WereCatf's post wasn't there when I first hit the comment button, but we're basically trying to make the same point.


Edited 2010-09-09 05:53 UTC

Reply Score: 2