Samsung Begins Mass Production of Industry's First 3D NAND Flash

← Back to Stories (view on slashdot.org)

Samsung Begins Mass Production of Industry's First 3D NAND Flash

Posted by Unknown on Tuesday August 6, 2013 @12:05AM from the stacked-bits dept.

Lucas123 writes "Samsung has announced it is mass producing the industry's first three-dimensional (3D) Vertical NAND (V-NAND) flash memory that breaks through current planar NAND scaling limits, offering gains in both density and non-volatile memory performance. The first iteration of the V-NAND is a 24-layer, 128Gbit chip that will eventually be used in embedded flash and solid-state drive applications, Samsung said. It provides 2 to 10 times higher reliability and twice the write performance of conventional 10nm-class floating gate NAND flash memory. Initial device capacities will range from 128GB to 1TB, 'depending on customer demand.' 'In the future, they could go considerably higher than that,' said Steve Weinger, director of NAND Marketing for Samsung Semiconductor. Samsung's process uses cell structure based on 3D Charge Trap Flash (CTF) technology and vertical interconnect process technology to link the 3D cell array. By applying the latter technologies, Samsung's 3D V-NAND can provide over twice the scaling of current 20nm-class planar NAND flash."

56 comments

Min score:

Reason:

Sort:

You'd think ComputerWorld would know better... by msauve · 2013-08-06 00:12 · Score: 1, Insightful

It's not 128 GB (gigabytes), it's 128 Gb (gigabits) of capacity.

"128 gigabit (Gb) density in a single chip..." - Samsung press release

--
"National Security is the chief cause of national insecurity." - Celine's First Law
1. Re:You'd think ComputerWorld would know better... by Anonymous Coward · 2013-08-06 00:19 · Score: 2, Informative
  
  128 gigabits per Chip and 128 gigabytes per device.
2. Re:You'd think ComputerWorld would know better... by gl4ss · 2013-08-06 00:55 · Score: 2
  
  chip size is different from device capacities(multiple chips, in device sizes from x to xx), one is talking about the other.
  
  --
  world was created 5 seconds before this post as it is.
3. Re:You'd think ComputerWorld would know better... by jones_supa · 2013-08-06 01:03 · Score: 2
  
  You'd think ComputerWorld would know better...
  It's not 128 GB (gigabytes), it's 128 Gb (gigabits) of capacity.
  "128 gigabit (Gb) density in a single chip..." - Samsung press release
  No no no. Please read carefully. The ComputerWorld article says that the final device ("once used to create embedded memory and solid-state drives") can have from 128GB to 1TB of storage. However the article states erroneously "Samsung's new V-NAND offers a 128 bit density in a single chip".
4. Re:You'd think ComputerWorld would know better... by Lucas123 · 2013-08-06 03:47 · Score: 1
  
  No. Actually the press release and the article state that it's gigabits.
still don't get why I'm supposed to be excited by Trepidity · 2013-08-06 00:20 · Score: 1

Slashdot seems very excited about Samsung NAND lately but I don't really get it. Is this really anything but an expected incremental improvement? Is there something I'm missing that makes this super-futuristic NAND OF THE FUTURE live up to the hype?

--
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
1. Re:still don't get why I'm supposed to be excited by hattig · 2013-08-06 00:21 · Score: 2
  
  It exists? It means that NAND can live another generation or two longer as a viable technology before the competitors (e.g., RRAM) are ready to step in.
2. Re:still don't get why I'm supposed to be excited by sosume · 2013-08-06 00:32 · Score: 4, Funny
  
  US President Barrack Obama has been quoted saying "since this NAND type RAM is such basic knowledge and essential for producing smartphones, Samsung should provide it for free to US companies, starting with Apple since they are already violating their precious design IP."
3. Re:still don't get why I'm supposed to be excited by Anonymous Coward · 2013-08-06 00:37 · Score: 3, Informative
  
  I'd say this is up there with perpendicular recording, http://en.wikipedia.org/wiki/Perpendicular_recording for hard disks
  As the summary suggests it enables higher densities and improved transfer rates as we bump up against space and power limits based on physical limitations of process engineering. The fact is that 10nm is bumping head long into some fundamental laws of physics and the ability to scale storage to the point where it reaches the densities of spinning platters is still a factor or two away is a huge market challenge to move Solid State Data Storage off the specialty and into the mainsteam standard. If we can get 4TB of flash storage in a 3.5" form factor for $250 bucks, it kills the spinning disk. This gets us two or three steps closer.
  And unlike the RRAM story, it's an announcement that it's being placed in mass production, not some lab where it was assembled by hand and "still 5-10 years away," like nearly every storage story we ever see.
4. Re:still don't get why I'm supposed to be excited by stewsters · 2013-08-06 00:44 · Score: 1
  
  Its only funny because its so true to character.
5. Re:still don't get why I'm supposed to be excited by dbIII · 2013-08-06 00:45 · Score: 2
  
  If it's got enough depth it can live up to the hype since the main cost for these things is mm squared area on a wafer. That zone refined silicon is stupidly expensive in terms of energy and thus heating costs.
6. Re:still don't get why I'm supposed to be excited by Anonymous Coward · 2013-08-06 01:06 · Score: 0
  
  Because incremental improvement (Moore's Law) doesn't just happen. It takes research, creativity, and new technologies. This is one of them, and its interesting. Maybe Slashdot is not the aggregater for you.
7. Re:still don't get why I'm supposed to be excited by TheSkepticalOptimist · 2013-08-06 01:15 · Score: 1
  
  Well then, I would say that smartphones have become such a fundamental human right that US companies like Apple should give away iPhones for free since Apple has violated so many human rights overseas.
  
  --
  I haven't thought of anything clever to put here, but then again most of you haven't either.
8. Re:still don't get why I'm supposed to be excited by Anonymous Coward · 2013-08-06 01:21 · Score: 0
  
  its THREE DEE!!!
9. Re:still don't get why I'm supposed to be excited by Anonymous Coward · 2013-08-06 01:22 · Score: 0
  
  Is there something I'm missing that makes this super-futuristic NAND OF THE FUTURE live up to the hype?
  The future doesn't happen by itself.
10. Re:still don't get why I'm supposed to be excited by Anonymous Coward · 2013-08-06 01:31 · Score: 0
  
  The fact is that 10nm is bumping head long into some fundamental laws of physics and the ability to scale storage to the point where it reaches the densities of spinning platters is still a factor or two away is a huge market challenge to move Solid State Data Storage off the specialty and into the mainsteam standard.
  I'm not sure I agree with the idea that large storage capacity necessarily is the mainstream for disks.
  From what I can tell from users outside the slashdot crowd the cheaper SSDs are already in the size range that works well for the mainstream usage. The problem is that the cheaper SSDs aren't cheap enough.
  Not that I complain, I don't mind the idea of a 4TB flash drive at all and this might very well kill off the spinning disks. I'm just nitpicking.
11. Re:still don't get why I'm supposed to be excited by timeOday · 2013-08-06 01:44 · Score: 1
  
  An increase in the exponent is pretty exciting to me! For decades we have been hearing about 3d holographic storage with incredible densities, or 3d processors, and it never appears beyond, say, dual layer DVDs. Granted, Intel Haswell also has 3d lithography, but it is still one layer of transistors, implemented by 2 or 3 layers of material, e.g. each transistor goes across, then up, then over. In contrast Samsung is using 24 cell layers and planning to build up from there. (Probably the heating issues are not far worse with logic gates and dram as opposed to flash).
12. Re:still don't get why I'm supposed to be excited by jeffmflanagan · 2013-08-06 02:05 · Score: 0
  
  Seems that way. All part of the dumbing down of Slashdot.
13. Re:still don't get why I'm supposed to be excited by Anonymous Coward · 2013-08-06 02:29 · Score: 0
  
  Still no word on the cycle life. Who cares if it gets great performance if it means I'm spending $400 every couple months.
14. Re:still don't get why I'm supposed to be excited by LordLimecat · 2013-08-06 02:33 · Score: 2
  
  Oh come on, Apple started this war. Probably not a good idea to go after the company providing so many of your components, but whatever.
15. Re:still don't get why I'm supposed to be excited by gl4ss · 2013-08-06 02:47 · Score: 1
  
  From what I can tell from users outside the slashdot crowd the cheaper SSDs are already in the size range that works well for the mainstream usage. The problem is that the cheaper SSDs aren't cheap enough.
  they're cheap enough, not big enough. you can't use them for archiving photos etc.. so if you go with a cheap small one(comparable to cheapest new hd pricing) you're going to have to go with a 1tb usb drive or something else to have the space to install games, store music etc..
  
  --
  world was created 5 seconds before this post as it is.
16. Re:still don't get why I'm supposed to be excited by Firethorn · 2013-08-06 03:01 · Score: 1
  
  Unless you're crazy with the photos or are a professional 'most' SSDs will be plenty large enough. It's Video (and games) that exceed most HD's.
  You're still going to see HD's in most things until Flash is cheaper for what the 'average' user needs/wants. One thing to remember is that at this point that there's a price floor for HD's. Such low end devices are more expensive per GB than the larger HDs because the mechanicals impose a certain minimum cost.
  We're getting closer, but for power users on a budget like me... They have a while to go.
  
  --
  I don't read AC A human right
17. Re:still don't get why I'm supposed to be excited by FloydTheDroid · 2013-08-06 03:39 · Score: 1
  
  Smartphone Wars - Apple is not the first company to sue nor are they always the first aggressor.
  This whole thing is just like an old-west bar-fight... if you don't punch someone in the face they're eventually going to break a chair over your back.
18. Re:still don't get why I'm supposed to be excited by Bengie · 2013-08-06 05:29 · Score: 1
  
  I think 256GB for $80 will make SSDs everywhere.
19. Re:still don't get why I'm supposed to be excited by LordLucless · 2013-08-06 13:56 · Score: 1
  
  Considering that all Samsung's other competitors have had to pay to use their patent, and Samsung has had to pay to use all the other RAND patents, yes, it's more than a little annoying to see Apple get a free pass for ignoring its licensing requirements all these years, just because it's the president's little love muffin.
  
  --
  Just because you're paranoid doesn't mean there isn't an invisible demon about to eat your face
You'd think smart asses on slashdot would know... by Anonymous Coward · 2013-08-06 00:30 · Score: 2, Insightful

If you lack the reading comprension skills and computer knowledge to correct someone, it makes you look twice as dumb as you think you are making the object of your correction look. Virtually no device is deployed with only one of these chips. Many are deployed in arrays of 8-32 or more. So if you have 8 bits, you have 1 byte. Consquently, if you have 8 chips of 128,000,000,000 bits you have 128,000,000,000 bytes. (Don't get into the pedantic Gibi versus Giga)
MODS this isn't insightful it is WRONG by Anonymous Coward · 2013-08-06 00:54 · Score: 0

The summary is right, they are indeed talking about 128GB to 1TB (BYTE) in initial *device capacity*
Let's... by Anonymous Coward · 2013-08-06 00:56 · Score: 0

Get perpendicular! Get perpe...
Sorry, wrong song.
This IS exciting but... by Anonymous Coward · 2013-08-06 01:07 · Score: 0

when will we get photonic circuits?
10D by EzInKy · 2013-08-06 01:26 · Score: 1

Why haven't we developed bits with ten or more states instead of two? This is something I've always wondered. Wouldn't that increase memory density and reduce the complexities involved with conversions from decimal to binary and back?

--
Time is what keeps everything from happening all at once.
1. Re:10D by Anonymous Coward · 2013-08-06 01:38 · Score: 0
  
  You're thinking of the old single-level cells, I present: MLC memory.
2. Re:10D by akh · 2013-08-06 01:46 · Score: 4, Informative
  
  We have, they're called multi-level cells. They improve memory density but at a cost of increased complexity, lower speed, higher susceptibility to noise, higher power consumption, and decreased lifetime. Decimal arithmetic was used on at least one early computer (ENIAC) but binary circuits were found to be much simpler to design and implement. The only modern non-binary digital computer that I'm aware of is the Soviet Setun that used ternary (tri-state) logic.
  
  --
  Accept Eris as your Fnord and personally sate her
3. Re:10D by Anonymous Coward · 2013-08-06 01:47 · Score: 0
  
  On a simplistic overview, MLC (multi level) nand already does this. Original flash (SLC - single level) stored a single bit in a capacitor. MLC stores it in the voltage of the capacitor. Determining the cutoff between one voltage and another is a lot harder than , is there a voltage or not. its also more prone to degredation since as it loses charge over time, it could be misread later as a different level.
4. Re:10D by EzInKy · 2013-08-06 01:50 · Score: 1
  
  From the link you provided, I believe I'm thinking more of a multi-level cell with ten states instead of four.
  
  --
  Time is what keeps everything from happening all at once.
5. Re:10D by u38cg · 2013-08-06 02:47 · Score: 1
  
  Because computer science and practicality.
  
  --
  [FUCK BETA]
6. Re:10D by loufoque · 2013-08-06 02:55 · Score: 1
  
  Because it's stupid.
7. Re:10D by Anonymous Coward · 2013-08-06 05:13 · Score: 0
  
  Binary logic is much easier to discern state[0,1] for processing.
8. Re:10D by Bengie · 2013-08-06 05:22 · Score: 1
  
  States are a power of two. Most MLC are 2 bits and TLC are 3 bits. MLC can be lots of bits, but the total number of states will be 2^(number of bits)
Hmm, Line noise? by Anonymous Coward · 2013-08-06 01:29 · Score: 0

Did the intertubes start leaking or something?
orly? by slashmydots · 2013-08-06 01:53 · Score: 2

"depending on customer demand."
I demand that they stop price fixing SSDs and RAM just because Windows 8 is selling like crap. Once a 500GB SSD is a reasonable price, I'll pick one up and they'll put Seagate and WD out of business completely. What the hell are they waiting for, sitting back and making ridiculous profits in the short term instead?
1. Re:orly? by MachineShedFred · 2013-08-06 02:24 · Score: 1
  
  Yeah, because DRAM producers like Samsung have never been convicted of price fixing before. Oh wait, they have. Multiple times.
  
  --
  Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
2. Re:orly? by Bengie · 2013-08-06 05:25 · Score: 1
  
  I remember the last time they got convicted of price fixing, everyone was selling memory at a loss. Competition was so high that they all decided they no longer wanted to sell for a loss, so they all agreed to increase prices at the same time.
  
  One of the few times I agreed with price fixing. Damn those prices were soooo cheap at the time.
Still good but. by Virtucon · 2013-08-06 02:20 · Score: 1

This will be a great breakthrough and with 24 times the density / sq. mm. maybe the 1gb/$1 SSD pricing wall will finally fall to around hard drives at around 22gb/$1 but I'm not throwing my hard drives away yet. One thing the article doesn't mention is cost, or anticipated quantity cost and if Samsung be license friendly, i.e., let another FAB produce the chips or will it be Samsung only territory for awhile?
I wonder if they'll let Apple buy the chips too?

--
Harrison's Postulate - "For every action there is an equal and opposite criticism"
1. Re:Still good but. by Anonymous Coward · 2013-08-06 02:59 · Score: 0
  
  Wake me up when I can buy it.
  They have been talking this tech for years. At least at this point they are sampling which is a good sign. But that means at least 8mo+ out. A 24x bump in size per pin package is most certainly a game changer though. However if they charge 24x what the normal size is... I would guess they are 2-4 months ahead of the competition on this.
  It is also interesting they are talking going backwards in flash tech to get more reliability and better write performance. Just stacking them out higher to keep up the density.
  Like how they compare it to 10nm. With the subtle implication they are using it (which they are not at this point in time).
2. Re:Still good but. by Anonymous Coward · 2013-08-06 03:15 · Score: 0
  
  Apple doesn't need that technology ; they'll just make small storage spaces hype !
3. Re:Still good but. by Anonymous Coward · 2013-08-06 03:35 · Score: 0
  
  The article conveniently glosses over the fact that the process size on these 24x stacks will be larger than the current 1D processes....
  These chips, when they see production, will only be a generational jump above the ~19nm flash we're seeing today.
4. Re:Still good but. by jon3k · 2013-08-06 06:16 · Score: 2
  
  I love the guys who come on slashdot to read bleeding edge tech news then complain that it's not a product in the marketplace yet. Cracks me up. You cold post this comment in basically everything on slashdot. You might as well go read engadget.
5. Re:Still good but. by Anonymous Coward · 2013-08-06 08:12 · Score: 0
  
  (orig ac here) Why? I used to get very excited about this stuff. But now? Not so much. They are talking Q3 2014 before we see product. Sample = someone may be able to get a hold of lab samples. Will prob slip to Q4 for some reason. That is a year out easy. They have been talking about stacking since early 2000 (see IBM). I have a *long* list of products that never saw the light of day. This one at least looks close. But I am in the 'when I can look it up on newegg/amazon/mouser' camp now. I can dream with the best of them but am not holding my breath for holographic memory :)
  'bleeding edge tech news' ... slashdot? You have not been reading SD for awhile have you?
Re:You'd think smart asses on slashdot would know. by TechyImmigrant · 2013-08-06 02:32 · Score: 1

> (Don't get into the pedantic Gibi versus Giga)
That's being pernickety, not being pedantic.
So am I.

--
I should use this sig to advertise my book ISBN-13 : 978-1501515132.
Re:You'd think smart asses on slashdot would know. by OlRickDawson · 2013-08-06 03:45 · Score: 1

I was going to be pedantic and point out that the word was "persnickety", but then I just found out that "pernickety" was an alternate spelling for the same word. I guess I'm a bad speller after all.

--
Ol' Rick Dawson had a farm EIEIO
Re:You'd think smart asses on slashdot would know. by msauve · 2013-08-06 03:59 · Score: 1

Flash is usually more than 1 bit wide, and there's no need for a minimum of 8. Good luck fitting 8 in a micro-SD card. If they were discussing actual end devices (as opposed to the flash chips, which are also devices), the numbers given are almost arbitrary and therefore meaningless - a device could be built with 1 or more chips.

--
"National Security is the chief cause of national insecurity." - Celine's First Law
Re:You'd think smart asses on slashdot would know. by TechyImmigrant · 2013-08-06 04:13 · Score: 1

I think, but do not know for sure, that 'persnickety' is more common the the United States of America, whereas 'pernickety' is what is spoken in British English. I certainly never heard 'persnickety' until I moved to the United States.
I could check, but I won't, because I cannot be arsed.

--
I should use this sig to advertise my book ISBN-13 : 978-1501515132.
Bad News for Desktop by Dajhan · 2013-08-06 07:05 · Score: 1

I think this is news for Desktop...imagine, tablets with 1TB or higher...smartphones with 1TB or higher...