Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/intel-exiting-3d-xpoint.10863/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Intel exiting 3D XPoint?

3D Xpoint was launched to great hype, but never met expectations. Now it looks like Intel is getting out. Thoughts?

https://www.nasdaq.com/press-releas...ash-technologies-joint-venture-20181018-01360

As a someone who has followed phase change memory technology for many years (including Crossbar and others) and a heavy holder of MU shares I believe from the outset Micron had a larger contribution and more sophisticated experience than Intel from the beginning. From what I could find out, this agreement was put in place at the outset and was considerably slanted in Micron's favor with Intel even having to provide financing for any cost involved in a split. I don't know the exact terms, but I feel because of Micron's experience in multi layer technologies applied to memory, they are the stronger partner in the deal. This is mostly educated conjecture on my part from reading everything I could get my hands on as far as 3D Xpoint. It looks like from everything out there in this area, this is the new memory technology that is the farthest along, with the most promise. Micron has stated full production to start in 2H 2019. So far there is nothing out there to compete. TSM has been working on Crossbar for over 7 years that I know of and this is what prompted me to meet Daniel Nenni to ask questions on the technology about five years ago at a Tanner EDA lunch and learn session. Intel has been marketing it under the Optane name without a lot of success, but from what I have read, it already has shown some great potential to upset the memory game. In Italy a research team has used it to cut down MRI times from forty five minutes to two for their application. I have heard of some other uses where it has shown similar potential. They are currently producing it at a joint fab in Utah. I'm just as curious as anyone else about this technology and feel if it works out to have true potential, Micron has a chance to rule the memory game. I also know TSM is using Crossbar for onboard memory for devices and the Chinese are also going to fab Crossbar technology in some form or the other. Phase change memory of all types, seems to be in its early stages with 3Dxpoint looks like it's holding the lead, but only time will tell. The only other technology is this area I know about is Everspin, but from studying the money flows, there isn't serious potential in this technology for a while. I'm a finance person more than tech guy, but money talks loudly and I follow the money trails.

investors.micron.com/releasedetail.cfm?ReleaseID=1079408&_ga=2.258856780.791891694.1539907594-391585422.1536760677
 
Last edited:
When I first read: "3D XPoint: 1000x faster than NAND, 1000x the endurance of NAND, and 10x denser than DRAM"

It made me think back to David Manners:

"For 40 years phase-change memory has been a useful tool for extracting cash from VCs, budgets from managements and investments from corporates. All of it has been spent to little or no avail earning it the Techno-Ponzi scheme soubriquet in the process."

Did you ever read "Non Volatile Memory" (now named: "efan999") poster on the eeTimes comments sections?

His very strong view -- was that PCM (I know, 3dxpoint isn't, apparently) is 100% Techno Ponzi, and would never be successfully commercialised in volume.
It was interesting reading the threads -- someone even found a Samsung phone with PCM in it at one point.

Turned out to be a mock-up, or a plant or similar.
 
It was announced last month that Intel would move its 3D XPoint development to its Rio Rancho (New Mexico) site. So maybe Intel and Micron are just working independently.
 
It was announced last month that Intel would move its 3D XPoint development to its Rio Rancho (New Mexico) site. So maybe Intel and Micron are just working independently.
Yup. They said after Gen2 they would work independently of each other.
 
It will be interesting to know who is going to own the IP and if Intel will continue on their own after the split. Any further info on this would be appreciated. From the conference call, it seem MU will own the IP.
 
The latest I have heard is Intel will move some production to China. Also, Samsung is looking to use everspin technology for onboard memory on some chips.
 
It will be interesting to know who is going to own the IP and if Intel will continue on their own after the split. Any further info on this would be appreciated. From the conference call, it seem MU will own the IP.

I would guess they cross-licensed already (up to this point). On the patents, it is assigned to one or the other.
 
At the time, my understanding was that Intel brought the technology to MU in the joint venture.
 
Optane is quite a unique product, I believe Intel has interest in continuing to explore its market potential.
I just never understood why Micron hasn't released its own Optane alternative.
 
Because Optane was disaster. Average performance (expect randomized read with extremely shallow queue which was questionable use case in today operating systems), high power consumption (bad for servers, laptops, mobile... and other power and TDP sensitive areas), questionable endurance (Intel claims high endurance but some say opposite), high price.... and capacity... They were brave enough to release 16GB SSD in 2017. 16GB hard drive was usable 15 to 20 years ago and today you hardly fit operating systems with bare minimum of software in it.

Only reason why people are buying it is because it is Intel and they made lot of pr around it. Micron did not released it probably because they don't have customers loyal enough to accept it.
 
Intel's Response: Micron’s Control of 3D XPoint Fab Will Not Disrupt Optane Roadmap

Intels statement:

Micron’s statement is a pre-announcement. They can’t officially make the call until January 1, 2019. The operation of the IMFT factory would not change until after the close of the call, which is at Intel’s discretion for up to one year. There is no near-term change to Intel’s plans in the coming quarters—this has been part of our planning for some time now. Intel has a number of manufacturing options available to us within the time window. We’ve been shipping a broad portfolio of Intel Optane technology products for over a year with a continually expanding product line. We will continue to lead the industry with this exciting new technology.
 
I mentioned this on another thread, but does anyone know what file system manages this resource? If you add a new memory category it will not go far without intelligent software to put it to the correct use. We can manage disk and DRAM through many vendors. We are getting up to speed on NAND with some new vendors and some old face lifts. But does anyone manage this or does it act enough like flash to fit under that storage management software?
 
For optimal value from Optane-like devices (byte-addressable, high performance) which you want to use as persistent storage (as opposed to, eg, using it as a paging store or a RAM disk) you need a file system (or equivalent like a database).

A traditional file system *will* work (ie emulate flash) but is sub-optimal for two reasons:
- it goes through the OS and layers of abstraction. For traditional storage the latency of these layers is negligible, but it's a significant cost when the distance to storage is a small multiple of the distance to DRAM.
- traditional file systems are designed to work with large chunks (eg 4kB size) and so do not take advantage of byte (or more realistically cache-line) addressing.

Work is being done on both these fronts. What is known publicly is Microsoft is working on a file system that is optimized for these characteristics like byte-addressable, and work is being done in Linux on how to make the control layer between app and storage as thin as possible.

A realistic analysis today, I think, would be that on the software side there are no lock-ins. I would expect that everyone you'd think has an interest (from Apple to Oracle to IBM) is working on one or both of the aspects I listed above; they just haven't said anything.

On the hardware side, hmm. Here's my analysis.
Intel, now in its decadent phase, is being run increasingly by a finance mindset (look for rent opportunities and exploit them as long as you can) rather than an engineering mindset (spread new tech as widely as possible because somehow it will take off and smart people can make money from that). So everything around Optane was designed to force enterprise customers to buy expensive special Xeons and expensive special Optane DIMMs, with no interest in the wider world beyond deep-pocketed enterprise.
Micron, still an engineering company, considered this an idiotic, short-sighted strategy, and so took back as much control as possible over the tech, and wants to sell it like any new tech you want to take off, at a range of price points, with huge ambitions that it spread everywhere (just like flash).

So where does that leave us? ARM v8.2 (2016) included the sort of cache control instructions needed to control persistent memory, and presumably these will trickle out over the next few years into ARM server and lower end cores. (Apple definitely uses some ARMv8.3 features, but I don't know for definite if anyone has implemented these cache control instructions yet.) Presumably at some point AMD and IBM will do likewise, but I have seen nothing from them yet.

So, for another year or three, likely Intel will have the game to itself but fairly soon (before 2025, maybe 2023) I expect a commodity spec for the DIMMs, Micron to supply them (with, who knows, maybe Nantero and MRAM and other weird options also providing DIMMs), all the relevant CPUs providing the instructions necessary to exploit this stuff, and all the relevant OSs providing their part.
 
name99, thanks for the detailed information. It sounds like the greatness in Intel died with Andy Grove. I guess from what I've heard, no one in the current executive team at Intel commands the respect or admiration of Andy Grove or deserves it. I hope I'm wrong and they have an up and comer that can restore their reputation. If there is, please let me know.
 
2023 is a long time to wait. Let's hope someone does a patch that works by 2020. From what ASML management said, various Chinese vendors should ramp in 2019 and I imagine may impact DRAM prices by 2020. That is the year MU management hopes to ramp too, so that timing would be handy. Does MU have good driver engineers?
 
Perlmutter was the real hero and the difference maker. If intel can pattern at 10nm once or twice they'll be fine. There's some chaos there and people are playing politics. I'm curious about the truth behind the leak that they'll discontinue 10nm. If it's true or infighting.
 
Back
Top