Subscribe to EDN

Numonyx looks to bright future for NOR Flash and phase-change memory

October 9, 2008

With the current focus of attention on NAND flash as the drama queen of non-volatile memory, there is a tendency to overlook the importance of NOR Flash. And in this oversight, one hardly hears any more about the best contender for next big thing in the memory arena, phase-change memory. But Numonyx, the non-volatile memory vendor formed by carving out pieces of Intel and ST Microelectronics, is bullish on NOR and has phase-change memory very much in its sights.

Glen Hawk, vice president and general manager at Numonyx, sees NOR moving beyond its traditional role in code and static-parameter storage. At one extreme, Hawk agrees with the emphasis that Spansion has put on NOR/NAND combinations as an alternative to the enormous power appetite of DRAM in server farms. But Hawk sees this specific application as just part of a larger picture.

"As we move to finer geometries, scaling NAND Flash is getting more and more problematic," Hawk said. "We are seeing issues such as dropping endurance and rising error rates. In many cases, vendors are concentrating on cost and simply telling customers that they will have to compensate for these issues in their systems designs. We have seen some systems that use twice the necessary amount of NAND Flash simply to overcome the problems with the chips."

Hawk says Numonyx prefers another approach—using alternative technologies to compensate for NAND’s scaling problems without imposing systems design changes on the user. "For example, there are lots of instances where judicious use of NOR arrays can compensate for NAND’s problems," Hawk said. But eventually, even with help multi-level NAND runs out of steam, as the available cell area simply becomes too small to support enough stored charge. At that point phase-change memory comes to the fore.

"Today, we are sampling a 128 Mbit phase-change part, and we are getting good feedback on it," Hawk said. "But frankly, it is a specialty device at this point. To understand the technology better we chose to do the chip in 90 nm, so it is a full node behind the current Flash devices. For most applications today, conventional NOR Flash is a better choice.

"But there are two points that will make phase-change memory vital in the future. First, scaling. By the time we get to the 45 nm node next year, I think we will see growing interest in phase-change. And by the 32 nm node, the case for phase-change will be compelling. At that point, we believe it will be the only technology capable of providing backward-compatible non-volatile devices.

"The second point is process compatibility. It is a real problem to integrate Flash with a CMOS logic process. But we have already established that the phase-change technology integrates very well with standard CMOS. And it offers significant performance advantages, such as the ability to execute code in place."

This creates the possibility for new architectural options on SoCs and microcontrollers that don’t carry the penalties of either embedded Flash or embedded DRAM. If Numonyx continues to be happy with their initial adventure into phase-change memory, watch this space.

Posted by Ron Wilson on October 9, 2008 | Comments (9)

November 9, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
realist commented:

For those who think they understand the "industry," the reason why PCM will not commercialize is because the companies who can best do so are already producing Flash. Product cannibalization is shooting yourself in the foot. Like the netbooks killing all other PCs.


November 2, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
PETE commented:

I Agree there seems to be a fear in the industry that phase change memory will be commercialized...lots of competitors could be put out of business..


October 28, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
Sergey Kostylev commented:

There are no doubts that chalcogenide alloys are the best candidates for the future universal memory. Will they always remain alloys of future depends on achieving structural and compositional reproducibility of devices in mass-production.


October 28, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
Guest commented:

I think most posters here have a bias for or against PCM and science has nothing to do with the statements presented.


October 13, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
Guest commented:

PCM is the most likely alternative NAND/NOR architecture below 32nm. The question is: will Numonyx last long enough to mass produce the technology at that node? IMHO, the answer is no, unless it does a JV with Hynix or Toshiba.


October 10, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
Guest commented:

There is absolutely no way on earth Numonyx will get phase change to be anywhere near the cost per bit of Flash at 45 nm and below. The once-simple process has become a nightmare of complexity with "microtrenches" and other major tweaks. There is growing evidence that they have already doomed this technology into the hell of "niche markets". Spin all you want boys but you are fooling only the desperate.


October 10, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
guest commented:

I am amazed that people in the industry still take these characters seriously. Numonyx should publicly disclose their preliminary spec sheets for the 128Mbit samples, so everybody can see whether the phase-change memory is "good" or whether it lags NOR and Flash so badly that it would never be commercialized in volume, not at 45nm and not at 32nm.


October 9, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
guest commented:

Scaling PCM is not without its problems either. The resistance can drift between accesses, so verification is needed for each read. That doesn't sound any more reliable than Flash or DRAM.


October 9, 2008
In response to: Numonyx looks to bright future for NOR Flash and phase-change memory
nvm.expert commented:

With ASP's of ~$1/GB, NAND bit consumption has gone through the roof while most NAND manufacturer?s profits have gone through the floor. You are right; using NAND has gotten to the point where embedded customers are spending too much time and hidden costs to make it work with nothing but rocky roads ahead as the industry wrestles with scalability. I can't wait for the "good old days" when we no longer have to worry about data retention, read disturb and endurance. The industry needs PCM NOW!

POST A COMMENT
Display Name
captcha

Before submitting this form, please type the characters displayed above. Note the letters are case sensitive:

Advertisement
Advertisement
Advertisement
About EDN   |   Site Map   |   Contact Us   |   Subscription   |   RSS
© 2012 UBM Electronics. All rights reserved.
Use of this Web site is subject to its Terms of Use | Privacy Policy

Please visit these other UBM Canon sites

UBM Canon | Design News | Test & Measurement World | Packaging Digest | EDN | Qmed | Pharmalive | Appliance Magazine | Plastics Today | Powder Bulk Solids | Canon Trade Shows