Melting Memory Chips In Mass Production 117
chill writes "Nature is reporting that 'South Korean manufacturer Samsung Electronics announced this week that it has begun mass production of a new kind of memory chip that stores information by melting and freezing tiny crystals. Known as phase-change memory (PCM), the idea was first proposed by physicists in the 1960s.' With transistor-equivalent cells only 20 nm wide, switching time is around 16 ns. The first target market is cell phones, but the companies behind the technology see applications in PCs, servers, and other devices as well."
awesome (Score:5, Interesting)
i've been waiting for pcram to show it's head in consumer electronics for a while now. it has the advantages of being hundreds of times faster than flash along with having at least ten times the write-cycle life. it could turn out to be the OLED to DRAM's LCD.
the main disadvantage is that it's rather heat-sensitive, since writing is accomplished by melting crystals with a low melting temperature.
Re: (Score:1)
Re:awesome (Score:5, Funny)
Yes, and they've genetically engineered tiny little sharks to swim around zapping the crystals on and off.
Re: (Score:1, Offtopic)
In the good old days, the shark reference would have been in a GNAA troll. And it would have been a First Post.
Re:awesome (Score:5, Informative)
The crystals are melted by passing a current through a BJT transistor. The heat given off melts the calcoginide(sp?) material. Reducing the current quickly causes it to freeze in an amorphous state, cooling it slowly produces a crystal. The resistance of the two phases is different, thus having memory.
Pros:
*Naturally rad hardened since it is a physical state, not a charge like flash and DRAM.
*Easy to erase in manufacture (the reflow temp is high enough to erase the whole memory)
*3D memory arrays are possible. The same material can be used (with metal) to make a transistor, thus you can make layers of arrays. Traditional flash is one layer deep as it requires doped Si for the transistor.
Cons:
*In prototypes we've seen cell phones erase themselves when left in a closed up black car on a black dash with a black interior on a hot Phoenix AZ summer day.
*you can't factory program the memory, it must be programmed after reflow onto the device. (flash can be ordered from the factory pre-programmed in large unit orders).
*Manufacturing defects have been an issue (bubbles in the calcoginide material.
I used to work for a company developing this stuff. We had prototype units in modified production cell phones as long as two years ago (when I left). Not sure if some of the cons have been fixed since then.
-nB
Re: (Score:2)
In prototypes we've seen cell phones erase themselves when left in a closed up black car on a black dash with a black interior on a hot Phoenix AZ summer day.
Did the rest of the phone survive?
Re: (Score:3, Informative)
without it's firmware it wouldn't boot, but yes, re-programming the phone brought it back to life.
I see the new temps Samsung has hit are high enough that this will not be a problem for them.
Re:awesome (Score:5, Informative)
Rather heat sensitive, in comparison to other technologies, but the critical temperature of GeTe grystals is around 446 Celcius.
At room temp this stuff is rhombohedral structure, at at 446C it changes to a cubic structure. The size of these tiny crystals is so small that this temperature is easily reached.
Note that no liquid phase is involved here, its simply changing from a glass structure to a crystal structure.
This 446C temperature is not likely to be reached in the absence of other heat related destructive events, regardless of how tight your jeans are.
Re: (Score:1)
Re:awesome (Score:5, Informative)
Note that no liquid phase is involved here, its simply changing from a glass structure to a crystal structure.
No. Glass is a form of liquid - it has no translational symmetry at the molecular level. Rhomboedral or cubic structures are not glasses because they do have translational symmetry and are therefore crystalline. The change is between two types of crystal lattice.
Re: (Score:2)
Right, my physics is long ago. Thank you.
I meant liquid in the common usage.
You reinforce my point, some here were speculating that the change was from solid to liquid, (in the usual sense of the word), which is not the case.
Re: (Score:2, Informative)
Re:awesome (Score:5, Informative)
Glass is an amorphous solid, not a liquid.
Re: (Score:1)
The facts are all good and well, but if one ignores all that and defines "amorphous solid" as a kind of liquid, then he's absolutely correct.
And that's the argument that he's making in an attempt to save face. The facts are so inconvenient, and embarrassing.
Re: (Score:1, Funny)
Really? Then why are old windows thinner at the top than the bottom?
Re: (Score:1)
Re:awesome (Score:4, Interesting)
Asking why old windows are thinner at the top is like asking why the sun moves across the sky to contradict the heliocentric view of the solar system, and so deserves a funny moderation.
Re: (Score:2)
Let me preface this by saying I don't believe that glass will flow noticeably over centuries.
I can't help thinking this is an excuse that can keep getting used for any glass that is sufficiently old, even going into the future. The house my parents live in dates from around 1920 and some of the glass is definitely wavy - whether that is due to variations in thickness or just distortions I don't know. Likewise I have no way of saying how old the glass is. Let'
Re: (Score:1)
Two things:
1. We've found old glass panes that are thicker at the top rather than the bottom
2. If the glass supposedly flows visibly x amount over hundreds of years, then all we need to do is have the precision to measure whatever fraction of x we have the time to let a pane of glass sit. With modern instruments we see no such flow.
Re: (Score:2)
Re: (Score:2)
In fact a liquid and an amorphous solid are not distinguishable, the difference is simply one of time and viscosity... pick a large or small enough time domain, and any liquid is an amorphous solid and vice versa.
Re: (Score:2, Funny)
Are you sure? It can get pretty hot in my jeans...
Re:awesome (Score:4, Interesting)
This 446C temperature is not likely to be reached in the absence of other heat related destructive events, regardless of how tight your jeans are.
Well... unless you count lithium battery fires... :P Although I suppose they'd count as 'destructive events'.
:) One thing I don't understand, though, is the write cycle life. Does the phase change substance gradually settle into a third state? Or does the heating mechanism 'wear out'?
Kidding aside, thanks for the rundown. It now makes sense.
Re: (Score:3, Informative)
Kidding aside, thanks for the rundown. It now makes sense. :) One thing I don't understand, though, is the write cycle life. Does the phase change substance gradually settle into a third state? Or does the heating mechanism 'wear out'?
According to this detailed paper [objective-analysis.com], it's being conservatively limited to a million writes for now because they have no real experience from which to determine true write lifetime. Right now, they don't know which will fail first: the phase-change material, or the BJT driver tra
Re: (Score:2)
it has the advantages of being hundreds of times faster than flash along with having at least ten times the write-cycle life
Hundreds of times faster? NOR flash is available with 50ns random access and 25ns serial access, so are you suggesting it can cope with 250ps serial access? That's 4GHz - I doubt it. 10 times faster is a bit more believable. It would be nice to see some links to actual specifications or performance data rather than just recycled marketing bumf, something which all the articles I've seen seem to lack.
This is clearly a vast improvement over flash, just because of the write speed and lack of complexity wit
Re: (Score:2)
And What About... (Score:4, Insightful)
The article looks very scarce on details other than the technology itself which, honestly, doesn't say much about the final product at all.
Re: (Score:2)
The article says there's a 128MB prototype now, and Samsung is coming out with a 512MB version. They acknowledge that this is small compared to present-day flash, but think that because of the problems inherent to making flash smaller that PCM will be the storage technology of the future.
You're on the ball...Re:And What About... (Score:2)
Samsung had a problem with K6X designated static ram chips. They would fail with the symptom being 'starts to work after a while' or 'starts to work when externally warmed'. They of course blamed the designer of the systems rather than offer to replace the defective parts. This failure happened across all packages.
It's a part you'll mostly find in embedded systems.
Re:And What About... (Score:5, Informative)
Lifetime - significantly better than Flash, 3 plus orders of magnitude more write/erase cycles before there's degradation.
Impact on overall computer heat & energy required to use - lower read power than Flash, no maintenance power (DRAM requires rewrite cycles as the bits decay)
Expected size - Initial model is 0.5 GB (512 MB) per chip. That's on a much larger fab process than current CPUs or DRAM though - expect that to increase rapidly once demand is established.
How about SSDs? (Score:2)
PCRAM's properties also make it sound interesting for archival storage. As long as you can keep the temperature at a sensible level it appears to be stable.
Re: (Score:2, Interesting)
Re:And What About... (Score:4, Informative)
Re: (Score:2)
Per-bit writing (or even per-word writing) makes a lot of other things interesting, given a sensible controller design. If you can address it just like RAM then operations like mmap() can be implemented by just mapping part of the device's address space into the process's address space. If it's sufficiently fast, then you don't need to cache it in RAM; the overhead from the extra copies will be more than the overhead from the slightly slower access. This, in turn, makes things like execute-in-place trivi
Re: (Score:2)
did you try reseting your memory? (Score:5, Funny)
help, my computer's frozen! nothings responding!
did you try reseting your memory?
how do I do that?
a few minutes with a hair dryer should do the trick.
Laugh, but.... (Score:3, Interesting)
Re: (Score:2)
Destroying these is no different than destroying a normal hard disk or SSD securely. Burn them or grind them into dust.
Re: (Score:2)
A hair dryer that can heat to 446*C? You wash your hair with tar?
Re: (Score:2, Funny)
Aye, lad! Tar and feathers! HAAAARRR!!
Re: (Score:2, Funny)
Aye, lad! Tar and feathers! HAAAARRR!!
So you'd say that this new technology is perfect for storing pirated media? Avast thar!
Re: (Score:2)
You take your hair dryer in the shower?
It would be more cool if it (Score:1)
Re: (Score:1)
Depending on the liquid, it might be hotter.
Re: (Score:2)
Re: (Score:1)
I was just making a stupid play on some words in your subject.
Re: (Score:2)
Re: (Score:2)
Many CPU's in the past had Halt and Catch Fire [allexperts.com] instructions.
Re: (Score:2)
hmmmm (Score:1, Interesting)
Re:hmmmm (Score:4, Informative)
Yes you are being paranoid. You already run a serious risk of losing all your data when you drop your Cell, so nothing changes here.
Other things will break first.
Re: (Score:2)
Am I missing something, or is the parent vastly exaggerating the risk of data loss here?
Re: (Score:2)
No not really. It is mostly luck and engineering that prevents the loss of data currently. Engineering makes storage devices a robust as possible but they can't proof against all angles. Perhaps the engineers layer the protection so that you have to transfer the shock through multiple layer of absorbtion to effect the core.
However a single shock to a particular point has caused me to lose the data on my phone (not the simcard though). That was an much older phone; I haven't been able to kill my current pho
Re: (Score:1)
I've dropped many a cellphone without data loss or breakage, I was just wondering if there was an increased risk with the particular workings of this technology.
Re: (Score:2)
See up-thread. The "melting" is at or around 446C.
So, still safe to bake it into a cake....
Re: (Score:3, Informative)
Apparently the crystal bond is very weak. I wouldn't want to lose my data because I dropped my cellphone. Perhaps I'm being paranoid?
Keep in mind that your intuitions about how substances will react to shocks are all based on objects of a particular size. Small objects are much less likely to break under a given amount of acceleration than large objects are. For an example of what I mean:
Re: (Score:2)
you mean E actually equals MCÂ ?
Re: (Score:2)
Lastly, drop a 747 off the same roof, and see how much damage it suffers
It still hasn't come down. :(
Re: (Score:2)
Experiments complete: the roof suffered no detectable damage in all three cases. Mind you, the 747 did leave scuff marks, but that was before dropping it.
Misleading (Score:2)
Re: (Score:1)
I do agree that even the weight of a cellphone won't be enough to damage the tiny cristals, but your car analogy was flawed
CDRW (Score:4, Interesting)
Re: (Score:1)
Re: (Score:2)
Speaking of which, how durable is this memory?
RAM operates a few orders of magnitude faster than CD-RW...how many writes can it handle?
for the paranoid among us (Score:2)
Re: (Score:2)
Thermite. Its the only way to be sure.
Re: (Score:3, Funny)
Frankly, I'm quite surprised I don't see 3.5 inch "thermite drives" being advertised all the time.
Re: (Score:2)
Bit excessive, don't you think?
We'll see if you still feel that way when they have you tied to a table giving you a red-hot-poker enema.
Oh yes... I will be vindicated.
Re: (Score:2)
Yes, it was definitely BS. Thermite doesn't detonate. It might deflagrate if you keep it in a sturdy enough container, but mostly it just combusts, rather like gunpowder, except that gunpowder doesn't evolve a pool of liquid iron.
Durability and Other Limitations (Score:5, Informative)
PCM is interesting stuff [wikipedia.org]. Here's some info:
Re: (Score:2)
Re: (Score:2)
Or flash it in-place. There are plenty of devices where the initial contents are loaded after the board is assembled. JTAG is the popular method - either via boundary scan, or debug mode, but more and more SoC's have a boot ROM that can be put in a special "download" mode. Short a few lines, hook up via
Re: (Score:2)
Or just use a socket and insert the chip later once things have cooled off, BIOS-chip style, like we've been doing for decades.
the 'weird force'? (Score:1, Funny)
is that the missing link between the electroweak force and gravity?
Re: (Score:2)
It'll be heat sensitive - weak crystal bonds will 'fail' if the module gets too hot. This also means interesting challenges around soldering.
I do not see the connection. You want to write the contents of the chip and THEN solder it? Moreover the phase change temperature is above 400*C, so far higher than typical soldering temperatures. If there is a problem with soldering at all, it would be the other way around (hot memory melting solder joints during writes). Heating the memory above the phase change temperature will erase its contents, but other components of a PC will not survive 400*C either, so it is not a big concern. Finally, there is no
Re: (Score:1)
PCM is interesting stuff [wikipedia.org]. Here's some info:
Unlikely. GeSbTe has a crystallization point of 400C. Well above typical soldering temperatures (230C-350C depending on process)
Better be careful... (Score:5, Funny)
If these things run too hot, you'll literally have vaporware.
Re: (Score:1)
Virtual vapo-ram becomes real vapo-rom
--
Why? To show the armadillos that it can be done.
Obligatory Strong Bad (Score:4, Funny)
And the Compy... just peed my carpet. [homestarrunner.com]
Tomorrow's hot technology... (Score:2)
Maybe this time?
And where are the other technologies that were going to displace the current leaders in the memory market?
Bubble memory?
FeRAM?
It would be nice to have another player in the game!
Re: (Score:2)
How hot? (Score:1)
Re: (Score:2)
Second post says the temperature is 446*C. However, since this temperature needs to be achieved in a very small fraction of the chip's volume when writing, I guess that writing all bits at once would translate to a heat-up of a few degrees when averaged over the whole die. This can be further reduced if the material surrounding the memory cells has a higher specific heat capacity than the crystals.
Re: (Score:1)
As you can see here [google.com], I have absolutely no idea what I'm doing but it's a really small number
Re: (Score:1)
Does this mean... (Score:4, Funny)
...that processors that support this type of memory will have to provide a Halt-and-Catch-Fire opcode?
Primary advantage, so far unmentioned... (Score:5, Interesting)
Perhaps the single most important advantage of PRAM has not even been mentioned yet. PRAM does not require the stupid block erase semantics of Flash--you can read or write as much or as little as you want, at whatever alignment, with no impact on performance. This also means that an SSD will be very simple, require no caches at all, and still have blazing fast write performance, even for synchronous writes.
PRAM will still require ECC algorithms, wear leveling, and bad block remapping, but on the spectrum of controller complexity, it is a lot closer to DRAM than Flash. (Incidentally, the same can be said of performance.) Reads and writes would still be buffered for queuing purposes, but this is very different from a cache; it is simply to allow requests to be pipelined from the storage controller.
Compared with the very simple constant time operations with PRAM, Flash is a dog. The controller must cache writes while it reads, erases, and otherwise shuffles blocks around. Moreover, as the controller operates with volatile memory, it must do this very slowly and carefully, or a power failure could severely corrupt the disk. (There are Flash SSDs with an onboard super capacitor to work around this, but they are obscenely expensive.)
Due to their inherent nature, even the best Flash SSDs have severely asymmetric read/write performance. The fact that only one company (Intel) has managed to produce a decent controller also betrays the immense complexity required to eek out even moderately acceptable random write performance. In my opinion, so called "SSDs" made with Flash don't even deserve that moniker, as they are more like a fast hard disk. (They still have a sort of geometry which constrains performance, and aren't anywhere near as fast as DRAM.)
PRAM will fix that, offering performance similar to a DRAM SSD. There are many companies banking on Phase-change RAM [wikipedia.org] to displace Flash memory, Intel included. The wikipedia page has a lot more info, but basically, PRAM is superior to Flash in every way, except that the data on a prewritten chip won't survive a trip through the wave soldering machine.
Re: (Score:2)
Oh, I don't know; in my head I hear a deep British voice singing, "I have to push the PRAM a lot"...
(It's only a model! Shh!)
Abbreviations!! (Score:1)
It's all rather confusing, really.
Re: (Score:2)
Say we polarize the crystal using a specified voltage in it's 'melted' phase... Then to retrieve the value or 'polarization' we have a photon (created with a photon emitter) go through it effectively polarizing it (the photon) and then using a photon detector converting the photon back into a voltage
Congratulations, you've reinvented the CDR!
Re: (Score:1)
Only if you're talking about a place on Mercury, or about to be engulfed in vast swathes of basaltic lava flows. Don't use this if you place your datacenter in Yellowstone Park, folks.
In both cases, you have more serious problems than losing the contents of your storage. Like being dead.
Re: (Score:1)
Re: (Score:2)
ATM I'm just glad people quit talking about using ATM for packet networks and nobody needs ATM for Type 1 fonts any more so I can go to the ATM to get some money without brainlock.
Re: (Score:1)
The article talks about 512MB chips with a 16ns switching time. Assuming you switch multiple cells in parallel, from a word if you're using it as a memory, more if used as storage, and it has a 16ns cell switching time. Writing bit by bit would allow a 7.4MB/s write speed. Writing 64-bits at a time ups this to nearly 500MB/s. It's not a replacement for DDR3, not even mobile DDR, but could be used as an intermediary memory level between RAM and storage, or the storage itself once capacities increase (when it