Matrix 3D memory is World's Smallest

nokiator writes "Most of the headlines about cool new high density memory technology are from DRAM or Flash manufacturers these days. Matrix Semiconductor, a small Silicon Valley start-up, broke the trend today and announced that the world's smallest 1-Gbit memory chip. Matrix's chip is an antifuse-based one-time programmable ROM. The total die area of the 1Gb chip is 31 square millimeters (smaller than the blue/red pills in the Matrix movie). Matrix claims that they can achieve this density through a proprietary 3D circuit technology that combines 150nm and 130nm process geometries. When Matrix moves to 90nm process technology, it should be possible to manufacture a 8Gb memory chip on a reasonable sized (i.e. cheap) die. There are many potential applications of this kind of low cost, very high density ROM technology, mostly in content distribution area. One 8Gb ROM chip would have sufficient storage capacity to store the contents of an entire movie using H.264 encoding."

Comments

Sweet!

[AKAImBatman]

I can think of one use right off the top of my head. Anyone remember the console design I suggested? Well, if these chips are cheap enough, it may actually make sense to go back to cartriges! Which means that copious quantities of graphics (including videos and prerecorded music) could be used in games for an inexpensive console system!

Anyone else have any good ideas for this chip?

P.S. Definition of an antifuse. Usually the type of thing you only learn about when you're playing with FPGAs, ASICs, and CPLDs. (The "history of programmable hardware" book that comes with Xilinx's Starter kit gives a good overview of the different technologies including antifuse chips.)

P.P.S. If I'm doing my math right, 1-GBit of memory is ~119 megabytes. 128 megabytes if you're calculating 1-GBit == 2^30.

Re:Sweet!

[kyle90]

How about an implantable encyclopedia for your brain? Would be better if it could be updated, of course, but you have to have certain trade-offs for it being so small.

Re:Sweet!

[thundercatslair]

I don't know about that. 128 megabytes is very small compared to the size of games today. Jade empire is over 5 GB, the cost would be unreal. I would love to see game companies move back to cartridges too, but unless there is a insane breakthrough in memory technology you will only see them in portables, even there they are dying.

Re:Sweet!

[poopdeville]

From TF blurb: The total die area of the 1Gb chip is 31 square millimeters... When Matrix moves to 90nm process technology, it should be possible to manufacture a 8Gb memory chip on a reasonable sized (i.e. cheap) die.

These things are tiny too. If you make a cartridge-sized one (even of the lower density variety) you can have a LOT of storage.

Re:Sweet!

[Short Circuit]

A CD/DVD's data area is around 10,000mm^2. That's enough area for 343 of these chips at their current size. 343*128MB=42.875GB

A better arrangement would be to make a 5" (127mm) square cartridge to fit in the same stackable region as a 5" circle. That gives you 520 of these chips. 520*128MB=65GB, which is better than Blu-Ray, and nowhere near as fragile.

And that's on their current process, which is apparently a blend of 130nm and 150nm. Wait until they shrink that down a bit.

Data density in the world has just gone up.

Re:Sweet!

[AKAImBatman]

I don't know about that. 128 megabytes is very small compared to the size of games today.

If you read the post I linked to, the idea was to bring back classic gaming at a low price point, not compete with today's games. I was originally thinking games along the lines of Duke Nukem II and Halloween Harry. But with this chip, we could jump all the way to Super Wing Commander! (With better voice acting, of course.)

Re:Sweet!

[Janitha]

Well you could leave a fraction of the ROM still blank and leave it for future changes. Just like current CD-Rs.

Re:Sweet!

[kesuki]

Well, if these chips are cheap enough, it may actually make sense to go back to cartriges!

For classic gaming maybe, for portable gaming sure. but you'll never get the price of solid state memory below the cost of optical storage. you can deal with the problems optical storage has currently by moving the laser beam with microscopic mirrors, rather than trying to spin the media. The problem with optical media is and was that they used design principals that work great for Magnetic media, and tried to pair that technology with optical storage. Since light can move exponetially faster (light can be moved to read 299,792,458 meters of data per second) than any physical device, it makes massively more sense to move the light, rather than the media or the laser. At current data densities.. that means the entire content of a DVD-rom would be read in 1/19933rds of a second. In other words, you'd have a 14,351,760x DVD-rom if instead of moving the DVD you move the laser light, and managed to move the laser light at the speed of light. There are of course scienitific limitations that prevent us from manipulating a beam of light in order so that it is redirected at the speed of light, but the theoretical limits of rotational speeds for DVD media are being reached. You can probabbly spin them faster than 16X, I seem to recall that at 1x a DVD-rom is moving the disc at the same rotational speed as 4x cdrom would be rotating and cd-roms got as high as 52x before cheaper media began fragmenting in peoples drives..

So what would you rather be capped with? 18x dvd-rom drives? or not have to worry about the engineering limitations until you can figure out a way to reach 14,351,760x?

Note: to those wondering, I based my calculations on the assumption that a dvd-rom has 4.7 billion bytes or 37.6 billion pits .4 micrometers apart, for a total length of 15,400 meters. I then used the knowledge that at 1x it takes 2 hours to read that distance, and calculated the speed rating based on that. I didn't check my math twice, so I might have made a miscalculation, but if I did someone will probably coreect me.

Re:Sweet!

[TechniMyoko]

These are the people Nintendo employes to make DS cards

Re:Sweet!

[Anonymous Coward]

Big gameboy games.

Re:Sweet!

[simcop2387]

there's a problem with moving mirrors like that you end up having to have such great control over the mirror e.g. towards the end of the device you'd have to have 1 billionth of a degree (probably wrong but not far off) of control with the motor moving the mirror also you wouldn't be able to catch or detect the responding array without having a mirror twice the distance away beaming it back to the origin (or have the sensor there, but you'd have to move that then with the mirror!)
quadurapling the total distance for each bit.

Re:Sweet!

[ScrewMaster]

Yes, but what I want to know is if this antifuse technology requires a dilithium matrix to regulate particle flow.

Re:Sweet!

[thundercatslair]

Well, you may be able to get 65GB in such a small area but what of the cost? I'm sure that optical storage will be much cheaper for a long time.

Re:Sweet!

[kesuki]

I realize there are engineering problems with moving the light instead of the media, one solution is to use a larger 'read' head like a CCD... also, centrifical tracks are less suited to moving the light, and disc shapes are less preferable than square or cubed shapes, and if you're using a mirror, a concave rather than a flat surface may solve some of the challenges.. you can also use refraction to split a beam and read multiple locations at once, etc etc... the point is that from an engineering standpoint it makes less sense to try and rotate the media than it does to figure out the best way to move the light. remember light can also be shifted via quantum events (ala a strong magnetic field) mirrors are just a lower power option..

I only did the math to demonstrate how we've crippled our optical read technology by trying to mimic hard drive read mechanisms.. Light is faster than anything out there, and if you're using light to read the data, you should be designing a way to use light to Seek the data too!

But thank you for reminding me that in fact the light has to travel a distance greater than the length of the pits. So for a 12 cm disc, a round trip would take at most 24 cm, it takes light roughly 1/100,000,000th of a second to make that round trip, but assuming a single non refacted beam you have to make that trip 37 billion times, the average distance would be more along the lines of 12 cm round trip, so it would take somewhere along the lines of 2 minutes resulting in an average read speed of 60X (but knowing the industry that 'average' read speed would be billed at it's 'peak' speed of 180X rather than at it's average speed of 60X ;)... Of course, by not moving the media, data density can be packed much closer, allowing smaller discs to have more data, and a laser beam can be refracted to read multiple sections of the disc, which if say, you write and read 16 bits of the disc at a time, you've increased the read/write speed by 16x and refraction is easier to calibrate when the laser and media are in a locked position, rather than a 'jittery' rotational one.

I liked 14 million x better than 160x, so I guess the laser is just going to have to be refracted to read 87,500 pits at a time ;)

Re:Sweet!

[EggyToast]

A lot of people must have forgotten about this.

About 2 years ago, Nintendo dumped a few million dollars into this company. Outside of being a useful technology on its own, it's basically being developed so that Nintendo can continue to use cartridges for their handheld consoles.

Remember how the DS has a maximum cartridge size of 1GBit? Yet nothing out now currently uses that? Hmm... awful convenient...

Anyway, that's exactly what it's being used for -- portable game consoles. No doubt it will work its way into other technology that values size and battery life, though.

Re:Sweet!

[Anonymous Coward]

I don't understand ANY of that.

Holy fucking GEEK Batman.

Re:Sweet!

[Anonymous Coward]

Fusable ROM chips should have a shelf life of thousands of years. They would be heat, cold, moisture, radiation, and chemical resistant.

This is the librarians wet dream.

An 8GB chip like this could store the entire Gutenberg collection and last for 1000 years with just a little care.

Imagine titanium time capsules that are lined with 10,000 of these and then filled to the top with amber and then the time capsule screwed shut. Then you bury one of these collections with an annual update every year in various places around the world.

--

I am thinking that 3D chips like this could be made in thousands of layers, like a cube. They don't draw that much power, a 1 is a short, a 0 is an open. I can see these circuits easily reaching Terabytes in capacity while staying below the size of a fingertip.

A solid state Write Once Read Many drive with hundreds of 1000 layer chips could be Petabyte's in size. The data in the library of congress could fit compressed on a single drive.

--

And if you needed to, you could blank the whole device by dumping a large charge into it to blow all the fuses open instantly.

--

Another cool thing that you could do with these chips is the fact that they can be accessed by 64bit machines like real memory. And you could wire it so that it could allow each layer to be accessed seperately and simultaneously by different processors.

Imagine having a 64 core system, 8 chips with 8 64 bit cores each all reading from different areas of 64 Terrabytes of ROM at the same time. 64TB of compressed text would be every book ever written, in every language and the entire compressed data could be read completely every minute.

--

Another great use for something like this would be data backup. Since these are write once, read many, you can just plug in the back up ROM and run a process to copy the data over from the old to the new, then take out either and you are done.

The same robots that are swapping out backup tapes in large computer systems could easily be changed over to read and write from cassetes that are filled with these TB of these chips. And the beauty of these chips over tape is that chips are random access at memory speeds.

--

Another fun thing is imagine you had a small rack in the computer that could accept 100 of these chips and the processor could access the programs and data on them at main memory speeds.

Installing an Encyclopedia including all software would be done this fast: "CLICK" And would be accessible continuously.

Since the software is ran from the memory in the ROM directly, you don't need to copy it to a hard drive.

--

You could have these in Palm computers, imagine being able to play video and sound on a palm computer for hours, since it takes very little power to read ROM. Since ROM is just like memory you could run programs from the chips directly, without having to install software.

Re:Sweet!

[jerde]

You're not trying to say that the speed of light has ANYTHING to do with the speed at which a beam of light can be swept across a medium, are you?

A point of light can be moved as fast or as slow as you want it to be. Aim a laser at the moon, now sweep it across the moon as fast as you want. Poof! The "spot" of light just moved across the moon at 18 times the speed of light... no problem.

No, the real limitation to the speed at which light can be moved along a medium has more to do with how long the spot of light must be focused on each point on the medium for enough photons to reflect back to be read. The faster you sweep your laser across a surface, the more dimly that surface is illuminated.

(At the limit, you're moving the spot of light faster than the rate photons are being emitted, although at 1 Watt that's something like 10^20 photons per second. If you need at least 10 photons per nanometer, say, you can do the math to find your maximum speed.)

Re:Sweet!

[TeknoHog]

A point of light can be moved as fast or as slow as you want it to be. Aim a laser at the moon, now sweep it across the moon as fast as you want. Poof! The "spot" of light just moved across the moon at 18 times the speed of light... no problem.

I was just about to say this, and let me add that this is consistent with Relativity, as long as you're not moving energy or information from one place to another faster than at c. In this case you're moving information/energy from the laser to the Moon, not between the points where the laser hits.

Re:Sweet!

[oever]

How about Knoppix chips for quickly booting, read-only computer systems. Of course, other OS vendors could do the same and sell their OS on a read-only USB stick.

Re:Sweet!

[Timberwolf0122]

If you cant spin the disc faster why not use a ballanced roptating mirror and rotate the beam. You could achive incredible speeds as the mirror would be v.small and light the hotizonalt positionn of the beam could be altered by using either a swash plate on the mirror (like a helicopter blades) or by angling the laser.

That said I do prefere the Idea of solid state chips more. how long befor we get 3D Ram??

Re:Sweet!

[Anonymous Coward]

Can't follow your line of argument. You suggest moving the light by moving tiny mirrors. Those are physical/mechanical and have to be moved themselves. How can you "move" a bundle of light so fast in a predefined curve (circle) without moving *any* mechanical device (being the source of light or a reflecting surface), as this would be the limiting factor, as you point out yourself.

Re:Sweet!

[MPHellwig]

ROM ... Read Only Memory?

Re:Sweet!

[golgotha007]

How about an implantable encyclopedia for your brain? Would be better if it could be updated, of course, but you have to have certain trade-offs for it being so small.

tell you what, I'll work on the storage medium, you work on the interface, deal?

Re:Sweet!

[schtum]

An 8GB chip like this could store the entire Gutenberg collection and last for 1000 years with just a little care.

Are you on crack?! The Police Academy series alone would take up three of these chips!

Re:Sweet!

[mr3038]

Two things:

1. The point where ray of light hits a surface can move faster than light. Information (light) goes from light source to surface or the other way around but it doesn't go from one surface location to another so you're not limited by the speed of light
2. You have to remember the little thing called wavelenght. You cannot read details smaller than the wavelength. The smaller the angle the ray of light hits the surface, the fussier the surface seems to be. We're going towards blueray laser systems because normal laser has too long wavelength even for next generation spinning disk technology.

If it wasn't the cost, one could make a pretty high speed optical reader by creating 12cm x 12cm CCD chip capable of reading every single pit in the surface of DVD disk at once. It would be just like a flat bed scanner, just faster. Scanning the whole disk would take 1/100th a second at most. Then it would be just matter of moving bits from the CCD chip to system memory fast enough. I won't even try to guess how much a 12cm x 12cm CCD chip with enough resolution to contact read a DVD disc would cost.

Re:Sweet!

[Kent Recal]

No, make a cube and paint the individual chips all funny.
Revenge of the rubics cube, encrypt your sensible data by physically scrambling it!

Re:Sweet!

[Short Circuit]

Scrabble?

"SETEC Astronomy"

Re:Sweet!

[stonecypher]

Anyone remember the console design [slashdot.org] I suggested? Well, if these chips are cheap enough, it may actually make sense to go back to cartriges!

I told you then and I'm telling you now: this is the brand of FROM that the Nintendo DS uses.

Re:Sweet!

[stonecypher]

For classic gaming maybe, for portable gaming sure. but you'll never get the price of solid state memory below the cost of optical storage.

You know, it's funny: I remember someone saying exactly the converse of this to me about fifteen years ago, when NeXT adopted the laserdisc as a standard storage mechanism for the Cube, back when ROM was considered cheap.

The older you get, the less likely you are to use the word never, especially in regards to the future.

Now, with an infinitely redundant power supply

[Trizor]

I can ramdisk the internet. I just need a warehouse!

Re:Now, with an infinitely redundant power supply

[AKAImBatman]

I can ramdisk the internet.

Except for the fact that this is ROM. It makes use of antifuse technolgy which works a lot like fuse technology. The idea in fuse technology is that you blow the pathways you don't want, thus creating the circut. With antifuse technology, the fuses don't normally conduct electricity so you have to blow the fuse to create pathways.

Info

Re:Now, with an infinitely redundant power supply

[Anonymous Coward]

Replace "wa" with "who" and you'll see what I see upon first reading your post. I think it may have actually been a bit funnier that way...

Re:Now, with an infinitely redundant power supply

[Trizor]

I RTFA'd AFTER posting, my bad. I thought I could have a first post, but I didn't notice having capslock on, and got owned.

Re:Now, with an infinitely redundant power supply

[patio11]

No problem, just go to the place you bought your infinitely redundant power supply and buy an infinite number of ROM chips -- all you need to do is keep a central log of the outdated ones and never try to access them again. Oh, don't forget to buy an upgrade to an infinitely large warehouse.

Oh good

[Anonymous Coward]

The total die area of the 1Gb chip is 31 square millimeters (smaller than the blue/red pills in the Matrix movie).

Just what I always wanted - another unit of measurement. How many millifootballfields is one blue pill? What can this chip hold in terms of LibrariesOfCongress-per-BluePill?

Re:Oh good

[Anonymous Coward]

.005 Libraries-per-Blue-Pill

and about .0073 Inverse Libraries-per-Red-Pill

It can't hold any Libraries on the white pill, but you won't care because it's crack.

Re:Oh good

[kf6auf]

31 square millimeters is between 5 and 6 mm on a side, or about half a centimeter (a quarter of an inch) on a side though to be more exact it's 0.31 cm^2 (go figure). Personally, I prefer to think of it as being 344 picoseconds^2 after setting c=1 and measuring distances in seconds.

Re:Oh good

[Nicholas Evans]

1 millifootballfield == .122173591659265 Space Shuttle's of Fuel.

Re:Oh good

[PaulBu]

Personally, I prefer to think of it as being 344 picoseconds^2 after setting c=1 and measuring distances in seconds.

At which Er? Si, GaAs, InP, Alumina, Quartz... or air? ... ;-)

Paul B.

Re:Oh good

[gibson042]

The only reason I read the comments on this article was to see which form the "slashdot units of measurement" post would take, and if it would make it to +5.

Re:Oh good

[Anonymous Coward]

Think we have a long way ahead to using metric units, when USians way of grasping the measurements goes by way of comparing with matrix blue/red pills. LOL!

Re:Oh good

[JaF893]

How many millifootballfields is one blue pill?
I've no idea but they have a mass of about 7.3 nanoelephants.

Re:Oh good

[ikkedus]

Don't know. Lets shoot another $125 million Mars orbiter into the sky and find out!

Re:Oh good

[Anonymous Coward]

I did exactly the same thing :)

Re:Oh good

[Tejin]

So, going on the assumption that you can walk at the speed of a bicycle, and carry 2 library of congresses worth of illegal music on five blue pills, how many football fields could you make it before you were sued?

Allow me to be the first to say...

[jm92956n]

Whoa!

Re:Allow me to be the first to say...

[Anonymous Coward]

Does anyone else find it hilarious that someone who posted a message with the subject "Allow me to be the first to say..." was modded redundant?

USB Linux-on-ROM

[Short Circuit]

I could do with seeing one of the fortold DVD-based Linux LiveCDs expanded even further and put on a read-only USB stick.

Oh, and it's OTP? You mean, like CD-Rs and DVD-Rs?

Comment removed

[account_deleted]

Comment removed based on user account deletion

Nice.......

[compmanio36]

Maybe this will help the emergence of solid state memory, as I find something like a Compact Flash card much more handy than a CD. I have had more DVD's that just wouldn't play because of the tiniest scratch on them. No, if there was a slightly more expensive, but much more reliable and robust form of memory storage, I would snatch it right up. Of course, I am waiting for my crystal-based isolinear memory chips that can hold gigaquads of data (whatever the hell a gigaquad is).

Re:Nice.......

[Anonymous Coward]

A gigaquad is one billion quads.

duh.

Re:Nice.......

[inject_hotmail.com]

...Of course, I am waiting for my crystal-based isolinear memory chips that can hold gigaquads of data (whatever the hell a gigaquad is)...

It is obviously one thousand times less than a terraquad...so...I'll wait for the terraquad model....or perhaps the petaquad or exaquad...yeah...I'm usually this finicky.

Inject.

Not just ROM's

[Transcendent]

...but could this be used for CPU on-die caches, or is it too slow/consumes too much power? I couldn't imagine even having 8MB of cache let alone 8GB. (Which will come to haunt me later like the ol' 640K quote).

Re:Not just ROM's

[WhatAmIDoingHere]

Unlike the 640k quote, you said this. Billy never said his "famous" line.

Re:Not just ROM's

[Short Circuit]

The technology dictates write-only. Even with a sliding write window ala multisession CDs, it wouldn't be useful as a cache.

Re:Not just ROM's

[Anonymous Coward]

Read only. Cache.

Re:Not just ROM's

[AKAImBatman]

Actually, it dictates more or less read-only memory. You can only write it if you have a programmer.

Re:Not just ROM's

[Short Circuit]

With all the power a CPU already uses, who'd notice a few watts more to burn ROM data? :)

Re:Not just ROM's

[AKAImBatman]

With all the power a CPU already uses, who'd notice a few watts more to burn ROM data? :)

You're worried about the CPU consumption? Pff! Have you seen GPU consumption lately? ;)

Seriously though, there's never been a case of a programmer being a standard piece of PC hardware. I seriously doubt this chip will change that.

Re:Not just ROM's

[antifoidulus]

A lot of Sun workstations already have 8 mb of L2 cache, for instance SunBlade 1000s. However, they like pretty much every other commercial chip in wide use(P4, Athlon, G4, G5 etc) all have 64k l1(32k data 32k instruction).
But more to the point, did you RTFA? This is anti-fuse technology, ie it cannot be re-written. I guess you could have certain chunks of data that you need to reference again and again that won't change, but for that limited use why would you ever muck up your architecture?

Re:Not just ROM's

[Short Circuit]

Yeah, I never understood that. If all you need to program is a specific model of chip, what's so difficult about embedding the burning equipment on the mainboard?

Granted, flash memory made PROM and EPROM BIOS obsolete. Heck...I sometimes wonder if they need the battery backup for anything more than the system clock these days.

Re:Not just ROM's

[AKAImBatman]

If all you need to program is a specific model of chip, what's so difficult about embedding the burning equipment on the mainboard?

A few things come to mind:

1. The process is *slow*. It's not exactly going to compete with CD/DVD burning. Or floppy disk writing in the old days. Or even worse, writing to tape!

2. In the old days, that WAS a lot of power to be pushing around a computer.

3. There's always the possibility that the chip could go bad during a write. If it does, any previous data you stored is now kaput.

4. Programmers are *bleeping* expensive!

5. Flash memory was around the same cost, and had none of these issues.

Re:Not just ROM's

[Transcendent]

Did you RTF Subject? And thank you for the sun link, I had no idea sun made such processors...

The point was asking if the hybrid scaling and segmented word line techniques could be applied to other memory types, unless there's an inherent physical impossibility (hence the question... I know they made ROMs... again, hence the question) in which volitale memory can't be made using the same techniques. It's a manufacturing process...

Why am I asking? I assumed the process for making the memory was pretty much the same since the number of transistors is pretty much equal (DRAM, not SRAM), it's just the final etching is different for the fuses (again, I'm just assuming). So can the process be applied to volatile memory, or does everyone just have smart-ass remarks to make and no actual knowledge?

Re:Not just ROM's

[HaydnH]

A lot of Sun workstations already have 8 mb of L2 cache, for instance SunBlade 1000s. However, they like pretty much every other commercial chip in wide use(P4, Athlon, G4, G5 etc) all have 64k l1(32k data 32k instruction).

They also make chips with 96K L1 cache (32K instruction & 64k data) with 16MB L2 cache for use in some of their servers, and remember they're RISC not CISC. For example, check out the data sheet for the E25K which can contain up to 72 of those beasties and 576G of RAM:

http://www.sun.com/servers/highend/sunfire_e25k/su nfire_e25k_datasheet.pdf (see page 2 under 'processor')

Anybody want a pocket calculator?

Haydn.

Re:Not just ROM's

[rpozz]

A few problems with it:

1) As you said, it's too slow. CPU cache is incredibly fast (because it doesn't need to be non-volatile).

2) You can only write it once.

3) All flash memory seems to have a limit to the number of writes before it screws up. Given how often the cache gets changed in a CPU, it really wouldn't last long.

Code named...

[Foxxz]

PRINGLES!

-Foxxz

Re:Code named...

[The-Perl-CD-Bookshel]

Bit (ok maybe a Gbit) off topic, but in memory of Mitch:

"I think Pringles initial intention was to make tennis balls. But on the day that the rubber was supposed to show up, a big truckload of potatoes arrived. But Pringles is a laid back company. They said 'Fuck it. Cut em up.'" Mitch Hedberg

Re:one movie?

[Short Circuit]

Try 10 * (1 byte / 8 bits), or one movie. GB is different from Gb.

Matrix memory?

[MutantHamster]

There is no 8Gb ROM chip.

Re:Matrix memory?

[Anonymous Coward]

You're right. It's currently at 1 Gb.

Density is fine, but speed ?

[karvind]

What is the latency of this memory module ?

Secondly it is antifuse-based one-time programmable ROM. It is NOT a flash which can be re-written 100,000 times. So it is more useful for storing application code but not for data storage etc.

Antifuse base memories are diode like and can be much smaller than regular FLASH memories. But these are inherently slower and also don't have any gain element (like transistor). This requires careful design to achieve good signal-to-noise ration for memory read operation

More aggressive 3D technology was demonstrated by IBM last year where they have circuits in 3D.

A startup R-cube logic is also designing 3D microprocessor where memory is put on top of the logic core to reduce latency.

Xanoptics is more into hybrid design (mixed analog, RF, optics) on a single footprint.

Re:Density is fine, but speed ?

[mcc]

Secondly it is antifuse-based one-time programmable ROM. It is NOT a flash which can be re-written 100,000 times. So it is more useful for storing application code but not for data storage etc.

Sounds great for something like a handheld video game system off the top of my head though. Handheld games are really hurting right now for need of some kind of compromise between hi-latency powerhungry high-capacity discs and low-latency power-cheap low-capacity ROM cartridges...

Revolution

[Anonymous Coward]

It'd be funny if Nintendo went back to cartidges because they could with this technology.

Re:Revolution

[stonecypher]

Yeah, um, this is what Nintendo currently uses in the DS, actually.

Some foresight required

[motorsabbath]

One 8Gb ROM chip would have sufficient storage capacity to store the contents of an entire movie using H.264 encoding.

Great, more disposable consumer things. There are many great uses for such a memory config, but the world does not need more disposable devices...

Re:Some foresight required

[Anonymous Coward]

One 8Gb ROM chip would have sufficient storage capacity to store the contents of an entire movie using H.264 encoding.

Great, more disposable consumer things. There are many great uses for such a memory config, but the world does not need more disposable devices...

Well, at least smaller disposable things might be considered a step in the right direction.

Of course, people don't generally just dispose of the things which store their movies. I mean, it's not like people just watch them once and then throw them out!

Re:Some foresight required

[motorsabbath]

heh. My nieces and nephews do similar things all the time. Just think about when they can get the most recent pile of pablum from Disney on a tiny, inexpensive, hand-held device to watch say 5 times and then smash/lose/etc...

Re:Some foresight required

[petermgreen]

we buy our movies on dvds which have the movie manufactured into them and are not rewritable...

if you are going to shell out the price of prerecorded media you are unlikely to want to record over it thus destroying most of your outlay (blank media is comparatively cheap in most cases by comparision)

Re:Some foresight required

[illumin8]

Great, more disposable consumer things. There are many great uses for such a memory config, but the world does not need more disposable devices...

Ok, mister enviro-conservo-nut-job step away from the tofu and put all animal friendly artifacts down! ;-)

Now that I have your attention, I would think, as an (assumed) environmentalist you would welcome the miniturization of potentially landfill-filling consumer goods. After all, isn't some thing smaller than a pill going to cause a much smaller overall environmental impact than, say, a DVD video disc or a VHS cassette tape?

Re:Some foresight required

[GreyedOut]

After all, isn't some thing smaller than a pill going to cause a much smaller overall environmental impact than, say, a DVD video disc or a VHS cassette tape?

The environmental impact of devices such as these should be measured not only in space they take in a landfill, but also on the manufacturing process and the resources needed to make them.

Re:Some foresight required

[motorsabbath]

Actually, it's neither the size nor the composition, but the disposable nature of the postulated devices. It's just plain wasteful and inefficient, that's all. Also, I got the phrase "consumer thing" from Zappa's "The Yellow Shark" - I'm not sure he was an "enviro-conservo-nut-job", although to the great unwashed I certainly am.

Re:Some foresight required

[Mac Degger]

Excuse me, mister I-didn't-really-think-this-one-through :P

How much energy do you think it takes to produce one of these ROM chips? Take into account shipping, handling, packaging...you know. Think of the production methods used to create these kinds of chiups: they're very messy, And all you can think of is "but it's miniature landfill"?

Grandparent poster's point is very valid: why consume those resources when you can just use some electrons/photons to transport the same data? Sure, this memorychip has it's place, but to say it's environmen-friendly because it's small is ignoring the poster's point that "disposable" is not very efficient.

Re:Some foresight required

[JamieKitson]

> Great, more disposable consumer things. If it does take off it will only replace currently disposable things, eg, cds, dvds, video tapes even (I know, I know, you can tape over the read-only lugs :)

This could actually be bad

[RhettLivingston]

ROMs can be very cheap. If they get to 8Gbit and then put it in a multichip stack to get to 4 or 8 GByte capacities, it could possibly give the movies on DVD industry a run for its money. The bad side of that is that we've been benefiting heavily from the demand that that industry has created is responsible for providing cheap RO and later WO and RW DVD drives for our PCs. The movie industry would love this format because the WO and RW versions would always be way more expensive than the RO version. The cost equation of copying would change dramatically.

Re:This could actually be bad

[evil-osm]

ROMs can be very cheap. If they get to 8Gbit and then put it in a multichip stack to get to 4 or 8 GByte capacities, it could possibly give the movies on DVD industry a run for its money. The bad side of that is that we've been benefiting heavily from the demand that that industry has created is responsible for providing cheap RO and later WO and RW DVD drives for our PCs. The movie industry would love this format because the WO and RW versions would always be way more expensive than the RO version. The cost equation of copying would change dramatically.

Same sentence as above, just replace ROM with DVD and DVD with VHS.

Re:This could actually be bad

[Anonymous Coward]

Actually, the write once version should cost approximately the same as the read only version, since they're in reality the same thing, only at different stages. The only major difference would be buying in bulk vs buying retail. I'd guess that ROM burners wouldn't be much more complex or expensive than current DVD burners were at first. Unless, of course, they package it in some complex DRM encumbered proprietary format.

Re:This could actually be bad

[RhettLivingston]

Not true unless this is not a true ROM technology. ROMs are typically not burned but instead made with a mask that creates them fully programmed from the start. The manufacturing cost, at least for runs large enough to easily offset the cost of the mask, is thus much lower than more complex one time programmable technologies which require more area per cell for some type of fuse mechanism and require an additional manufacturing step to blow all of the fuses.

Re:one movie?

[EightMillion]

8 Gigabit(Gb) == 1 Gigabyte(GB)
Check your math.

H.264

[after]

I didn't know what it was right away, so ...

H.264, or MPEG-4 Part 10, is a high compression digital video codec standard written by the ITU-T Video Coding Experts Group (VCEG) together with the ISO/IEC Moving Picture Experts Group (MPEG) as the product of a collective partnership effort known as the Joint Video Team (JVT). The ITU-T H.264 standard and the ISO/IEC MPEG-4 Part 10 standard (formally, ISO/IEC 14496-10) are technically identical, and the technology is also known as AVC, for Advanced Video Coding. The final drafting work on the first version of the standard was completed in May of 2003.

Amusing

[StarManta.Mini]

There's something inherently amusing about seeing this:
Allow me to be the first to say... (Score:0, Redundant)

Re:Amusing

[Koiu Lpoi]

Allow me to be the second to say... (Score:0, Redundant)

Re:one movie?

[compmanio36]

Thank you! I am always so damn confused about the difference between the two, but now I remember, 8 bits = 1 byte, so naturally, 8 Gigabits = 1 Gigabyte. Duh.

Some foreskin required

[Anonymous Coward]

"There are many great uses for such a memory config, but the world does not need more disposable devices..."

REUSABLE CONDOMS!

Re:one movie?

[d3am0n]

well...what the hell is so special about this thing then, I've already got a USB plugin memory stick that's teeny tiny and holds 1GB

Now, all we need are 3D Processors and RAM

[Xeroc]

It sounds like to me this 3D Memory construction is vastly improving the density, now all we need are 3D-Constructed processors! They use vertically and horizontally stacked chips to multiply the processing capability.

Also, if we could only get this in RAM! I'm looking for an upgrade, and my computer case is only so big!

Yes, for some reason, people do seem to mix up the bits and bytes, for example: Most file sizes are in bytes, to make them seem smaller, and connection speeds are in bits, to make them seem faster!*

*Actually, this probably isn't the "official" reason, but it makes sense!

Re:Now, all we need are 3D Processors and RAM

[mrmagos]

Files sizes are measured in bytes out of tradition, since a byte was originally the smallest unit of useful data (think ASCII and teletypes).
Connection speeds are usually measured in bits (actually bits/second) because they tend to be serial.

Fair assumption, though.

Re:Now, all we need are 3D Processors and RAM

[koko775]

Hmm. The problem is, though, 3D processors, unless redesigned, would have a far greater area:surface area ratio, which is bad, since the chip generates heat, and the surface area provides for its dissapation. I'm not saying 3D processors are impossible, it just adds a whole dimension to creating processors, as cooling has just gotten a lot harder. Also, antifuse won't work as RAM, unless you run Windows, in which case your computer, frozen forever, would work as expected (just kidding -- antifuse is write-once).

Re:Now, all we need are 3D Processors and RAM

[Anonymous Coward]

Create the processor with dimples. Choose a geometry to maximize surface area.

Re:Now, all we need are 3D Processors and RAM

[springbox]

I think it's because engineers are used to dealing with bit streams and most computer users are used to dealing with things in bytes. These units make it pratical for both sides, but it's too bad that a lot of programs fail to do the bits to bytes conversion when displaying speed.

Re:Now, all we need are 3D Processors and RAM

[Anonymous Coward]

I'm not saying 3D processors are impossible, it just adds a whole dimension

*Rim shot*

Not fast...slow

[WindozeSux]

Regardless of what the article may say about its speed, it is actually slower than everything because it will often make your computer do the "Matrix Effect".(You know when the people go real slow to avoid bullets)

Re:Not fast...slow

[Anonymous Coward]

With Bullet time, you are actually FASTER than the environment around you.

Blu-ray

[Anonymous Coward]

By the time these chips get anywhere near consumer production people will be buying films on HD-DVD and Blu-ray, which are 25-or-so gigs and use more efficent codecs than DVD*. Fitting HD movies in a 1cm^2 chip isn't going to be pretty.

*though maybe hobbyist codecs will be out that make divx look like MPEG 2 - even so, given compression by a factor of 5 you still can't squeeze a film to a gigabyte.

Re:Blu-ray

[robertca]

Matrix has already been mass-producing and selling their previous generation of chips, as described in numerous articles on the web and the Matrix website. From what I remember reading in the articles, the 1Gb part will be volume-ramped by the end of this year. On the other hand, there's no chance in hell that the Blu-ray/HD-DVD saga will settle down by the end of this year.

Re:one movie?

[Anonymous Coward]

This is 31 square mm, one square cm has 100 square mm. This is a chip, so maybe many cores can be stacked? A chip's thickness is 90% packaging.
I don't know about you, but I have never seen anything that small.

Long time in the making...but worth it...

[robertca]

Great to see the Matrix Semi news on Slashdot! I was one of the early employees (but have since left), so it's cool to see something that I worked on coming to fruition.

Earlier posters were correct in stating that it's not a complete replacement for flash (yet?) but there are still many very cool potential applications: Game cartridges (much faster access time than CDs/DVDs), toys (i.e. a supercharged Furby with a massive vocabulary), replacement for CDs/DVDs, archival digital "film", etc.

I really like the idea of a kiosk that houses blank Matrix 3DM cards and loads of digital content. You could walk up to the kiosk and buy a game/software/movie/album/book, have it programmed right then and there, and walk away with your customized content in a few minutes. These kiosks could be everywhere...gas stations, grocery stores, etc. Extremely convenient for consumers, plus it would seriously cut down the overhead for retailers since they wouldn't need to keep inventory or have huge stores to house thousands of DVDs, etc.

Re:Long time in the making...but worth it...

[Anonymous Coward]

plus it would seriously cut down the overhead for retailers since they wouldn't need to keep inventory or have huge stores to house thousands of DVDs, etc.

Or employ anyone related to such activities! GENIUS MAN YOU ARE NOW CFO

Re:Long time in the making...but worth it...

[PrivateDonut]

couldn't, by the same ticket, many stores have a machine that burns the dvd, then prints the image onto it. While doing that, it would make the case, by printing the slip of paper that goes in it, then printing the books/other pieces of crap they put in the case, then sticking all of these into a generic case?

Re:Long time in the making...but worth it...

[Dwedit]

The huge problem with that model is massive piracy. The people who own one of those machines can empty out the money any time they want, and print unlimited copies to sell at whatever price they desire without paying back the rightsholders.

Unless you have some way of making it completely secure, authenticated, and with no removable money, you'll have another Nintendo FDS Disk Writer fiasco.

Re:Long time in the making...but worth it...

[Anonymous Coward]

"Game cartridges (much faster access time than CDs/DVDs), toys (i.e. a supercharged Furby with a massive vocabulary), replacement for CDs/DVDs, archival digital "film", etc."

Why would manufactures want to base their products on a propriatary technology, when generic technology (flash) is more flexible? The manufacture can just disconnect the write_enable pin, and the flash becomes read-only.

I still fail to see what unique advantage outweighs all the shortcomings.

Re:Long time in the making...but worth it...

[robertca]

Matrix 3D is a fraction of the cost of flash. Also, while the semiconductor structure itself is proprietary (and heavily patented), the actual chip can be used in any sort of form factor, like SD, CompactFlash, etc. In other words, it can look exactly like your run-of-the-mill Sandisk flash card, but will function as a write-once media. The advantage of putting it into a widely accepted formfactor, like SD, is that it can then be used in any device that accepts SD cards at a fraction of the cost of flash...

Re:Long time in the making...but worth it...

[robertca]

It's better to have solid state memory (like Matrix 3DM) vs. a DVD for a variety of reasons... 1) More durable. 2) Potentially smaller and more versatile form factor 3) Significantly less power to read the media since it only involves electricity, not a motor and laser. 4) More durable (or did I mention that already?:-)

Re:Long time in the making...but worth it...

[Anonymous Coward]

I must be not getting something. How is that different from DVDs?

Re:one movie?

[Infinite Entropy]

I've seen 8 gig USB memory sticks. Breathtakingly expensive too!

In other news...

[Staats]

...Nintendo Revolution's cartridge based format is announced!

I see we have a new unit of measurement

[RichardX]

For data storage there's the good old Library of Congress (just how many LoC's DOES this store anyway?)
and for physical dimensions we now have "matrix red/blue pills". Interestingly, my mobile phone is exactly 67.4 Matrix Pills in total area.

Re:My thoughts

[Anonymous Coward]

Posting anonymously for obveous reasons.

YOU FAIL IT!

Jack in

[Digital_Mercenary]

Just let me know when the "brain-stem jack mod" is available. I could always use the extra memory.

Re:Jack in

[Kehvarl]

When you plug into a computer or peripheral, it's "jacking in". However, adding one of these for expanded memory or readily available content (information, basic skillsets, etc) is generally referred to as "chipping" or being "chipped".

Why, oh why do I know this stuff, and can I ever live down the fact that I admitted to knowing it?

Re:Jack in

[Digital_Mercenary]

I stand corrected... Thank you... ;)

Gb somewhat aloof measurement?

[gumpish]

I know that the bit is the atomic unit of measure when it comes to data storage and transmission, but sometimes I really wish everyone would stick to bytes.

When I see 1 Gb I have to think for a second to get to 128 MB.

Re:Gb somewhat aloof measurement?

[Anonymous Coward]

Nonsense! Memory chips are ALWAYS measured in bits!

Bytes are what do not make sense. The real consumers of these chips---the engineers who design around them---do not think in terms of bytes. They think in terms of bits or oddly-sized cache lines.

So get with the program. Bytes are such an arbitrary '70s concept. What's so magic about groups of 8 anyway?

Re:Gb somewhat aloof measurement?

[kahei]

Bytes are not always the same size; bits are.

Re:Gb somewhat aloof measurement?

[stonecypher]

1) Bytes vary in size, whereas bits do not.
2) Not all data streams are byte or even fixed-unit oriented.
3) It's really a remnant from baud vs bps, where baud was once a useful measurement but became less and then not so.

31 mm movies

[Anonymous Coward]

GUEST: Do you mind if I have some more m&m's?
HOST: What m&m's? I don't have any m&m's.
GUEST: In the bowl... on top of the tv...?
HOST: Aaaaaaargh! That's my movie collection, nimrod!

Smallest memory

[Junior J. Junior III]

And here, I thought that I had the -- wait, what whas I talking about?

Better Not Skip

[knapper_tech]

Holy shit! do DVD's ever suck!!!

You have to think out of the (X)box

[tepples]

128 megabytes is very small compared to the size of games today.

GBA and Nintendo DS games are no bigger than 32 MBytes, and the UMD dumps of PSP games are about 128 to 512 MBytes.

Re:You have to think out of the (X)box

[Hubertus_BigenD]

I believe this is the company that produces the DS game cards. If I remember correctly nintendo invested about $100 million in this company

YAY OS on a ROM - retro revival

[zenst]

Well with these size's we could be seeing OS's distributed on a ROM soon. Great anti piracy, well least add's another level that will remove a high percentage of copies out there. Remember nothing is perfect.

Roll on 3D holographic cubes for ROM storage, ready by laser :>

I'll skip!

[Anonymous Coward]

Then you see Intel patent 3D-Heat - you'll only need a bargain 2nd hand nuclear reactor cooling tower or 3 per cpu core.

Nintendo DS

[tepples]

Sounds great for something like a handheld video game system off the top of my head though.

Nintendo has already signed up with these guys to make PROM chips for Nintendo DS and GBA2 Game Paks.

Matrix memory is for DS games

[tepples]

You're on the right track, given this press release, but you misspelled "Nintendo DS".

Speech doesn't need a lot of storage

[tepples]

toys (i.e. a supercharged Furby with a massive vocabulary)

At some point, wouldn't it be cheaper to put a decent CPU in a talking toy to do formant based speech synthesis? Then each phrase could be stored as a simple string of X-SAMPA phonemes: [ka:'?EI.teI] for "I'm hungry". It wouldn't take an expensive CPU, as the Commodore 64 could do real-time speech on a 1 MHz 6502 CPU.

Heat

[tepples]

This is a chip, so maybe many cores can be stacked? A chip's thickness is 90% packaging.

And a lot of that packaging is a heat spreader. How would you efficiently move heat away from a cube of cores?

Sounds good for distribution

[jpardey]

Who needs a CD or DVD when you have a high density ROM chip? I don't see how access speed could be slower than an optical device, but I am no technician. Is this the "CD killer" that we have been waiting for?

Re:Sounds good for distribution

[Anonymous Coward]

It might not be as popular if it's only going to be read only.

Matrix 3D? Where's the originality?

[PocketPick]

Remeber when companies used to pick names that meant something or had some relation to whom or what was part of the company? Matrix 3D? In a world where companies pay thousands of dollars to come up with names for thier brands, I could of came up with that same name for a free lunch, $50 and two 6 packs of Mountain Dew.

Honestly. Do these marketing really come up with names or just scan a thesaurus and cross reference names with a trademark database. TorusTech Software Tools, MagicPrints Technology, SmartRights Consulting. Come on! I came up with these lame names in 5 minutes. Is that all it takes?

Re:Matrix 3D? Where's the originality?

[robertca]

When the name originated in '99, Matrix consisted mostly of engineers. At the time, there was no budget to pay someone to come up with a name, so the engineers were assigned the task of suggesting names that would somehow reflect the technology. Sadly, none of my suggestions won, and I missed my chance to leave a legacy there. The winning suggestion came from one of the process engineers, who, as far as I can recall, received absolutely no compensation for it. So, there you have it: The name describes the actual structure, was coined by a process engineer, and cost the company precisely $0 to come up with.

Re:Matrix 3D? Where's the originality?

[Hoi Polloi]

Oblig Simpsons Ref:

HOMER
This industry moves so fast it's really hard to tell. That's why I need a name that's cutting-edge, like CutCo, EdgeCom, InterSlice... come on, Marge, you're good at these! Help me out!

MARGE
How about... CompuGlobalHyperMegaNet?

Seriously though, they use consultants now to come up with these awful generic names for companies. It isn't much different than how they come up with names for new drugs. They just use a list of prefixes and suffixes that vaugly sound like the product or some image they want to create. I'm sure someone has come up with a program to crank out lists of possible names and cross reference them against trademark lists.

In the meantime try these: Cyber*, *.com, Dyna*, *dyne, *sure, *smart, *tech, Hyper*

Three letter acronyms are also a good choice.

I wonder if any actually uses a human name for a company anymore like Ford or SS Pierce.

Re:one movie?

[Anonymous Coward]

supposedly they can make em cheap, and like most flash and ram today, there are several of these 'chips' that make up a module. i.e. if they are cheap enough stacking a bunch together in a small space could yield enormous densities and low cost... could being the operative word.

Distributing linux distributions

[paul.schulz]

Set it up in a USBkey formfactor which
can be used to boot.

It would cost even less to send out Ubuntu CD's!

Good Gawd!

[Dark Coder]

If we approach 4GHz bus rate, thar be a core meltdown!

DL Blueray = 50GB, so no biggie

[Anonymous Coward]

dual layer blue.. ah, just read the heading.

Re:one movie?

[A. Rimmer]

"31 square mm......I don't know about you, but I have never seen anything that small"

GAAAHHHHH!!!!

Everything you've ever seen has been smaller than that.

smegging earth

Re:one movie?

[Anonymous Coward]

if you'd actually read the topic you will see that he specifies the H.264 codec which according to apple "H.264 delivers the same quality as MPEG-2 at a third to half the data rate and up to four times the frame size of MPEG-4 Part 2 at the same data rate."

Solid-state ROM burners in the future?

[Gactaculon]

If the cost can come down enough, these chips would certainly be useful for content distribution, as has been said. I'm left wondering, though, how tough it would be to create cheap, small ROM burners for use with PCs. Most people I know really like solid-state memory stick devices, but they see limited use due to their expense and small capacity.

What would be cooler than being able to throw a 25-cent, 1GB+, SmartMedia sized card into your computer and just burning it like you would a CD-R? I'd imagine that this process could become very, very fast, given that there are no moving parts and that the ROM could even be written to in multiple places at the same time, theoretically.

Coincidentally enough, when I think of what one of these burners would look like, designed for current computers, it would seem exactly like the removable storage drives on the Nebudchenezzar in the first Matrix movie...

Re:Solid-state ROM burners in the future?

[Anonymous Coward]

That might only happen if casing can be made that is cheap and doesn't have a lot to it. But if this did happen it would certainly replace CD-Rs for me. That is assuming it would be reliable.

Motherboards

[charlie763]

With the small physical size and large memory capacity might we eventually see a motherboard manufacturer shipping a Linux distro integrated into the motherboard? Not necessarily for the sake of using as you default OS, but as a distro with a full set of diagnostic tools. In any event, this sort of crap is way cool!

Re:Motherboards

[Lonewolf666]

Or a BIOS chip where a new address range is programmed for each update, with the option to revert to an older version.
You'd be limited to a certain number of BIOS upgrades (10? 100?), but if the upgrade fails, you could get back to the last version.

Matrix Memory?

[diablobsb]

Whoa!
Dejavu...

Re:Sweet! - you;re forgetting one thing

[lcsjk]

The problem is not the speed of light's round trip. The problem is that to move a beam of light, the light controller must sense position, tell the controller to move the light and then tell it when to stop and where it is at the time. Even if you are traveling along a straight line, the feedback mechanism still has to be there. The speed of the control electronics will be much slower than the speed of light traveling from a source to the reflections surface and back to the sensor. The speed of the electronics will be much faster than any type of mirror control available, including the microscale mirrors. And as the previous poster mentioned, you also have to position a sensor along with positioning the light source. This requires a lot of analog and digital circuitry and the speed/time envolved

Perhaps the biggest advantage would be the ability to treat the array as random access instead of rotational access with the inherent rotational delay.

is size the only selling point?

[chipace]

It's nice to be the physically smallest rom, but what about power consumption (static and dynamic)? Also, what about access time (usually you have to balance power consumption and access time... you get one at the expense of the other).

Flash die can be die-stacked to increase the storage capacity... using die from the same wafer. If you want to die-stack rom die, all the die will be from different wafers (extra mask layers are required for each unique rom die). Flash wafers will be cheaper to produce because they will have a much greater volume.

I'm sorry but flash looks like a much more flexibile and cheaper solution (being relatively larger in size).

Re:is size the only selling point?

[eluusive]

It's nice to be the physically smallest rom, but what about power consumption (static and dynamic)? Also, what about access time (usually you have to balance power consumption and access time... you get one at the expense of the other).

This is ROM as you astutely pointed out. ROM is not classified as "static" or "dynamic." It's Read Only Memory. As the data is permanently(non-volatile) stored it's power requirements should be very small.

Re:is size the only selling point?

[chipace]

"ROM is not classified as "static" or "dynamic.""

The data storage itself doesn't consume power, but the address_decode/read_mux is transistor based and comsumes static and dynamic power. The more storage you want to access, the larger the read_mux that is required to decode each unique address.

You are absolutely right that ROM storage does not consume any power... you just can't see the data without consuming power.

The threshold voltages for the process need to be low enough for the read_mux transistors to toggle quickly, but not too low that it leaks more power than you can spare.

Cool

[midnight2038]

Another great technology to be attacked by the RIAA, first in name "3D Matrix", and then for it's use!

Re:one movie?

[Anonymous Coward]

Your USB memory stick thingie is multiple chips in one package, not one chip! THAT's what's special!

Videogames, anyone?

[Esion Modnar]

Imagine going back to cartridge-based systems, instead of DVD-based?

3D memory is much easier than 3D processors

[reversible physicist]

A problem with 3D processors is that they contain a lot of active circuitry, so it's hard to get the heat out (surface to volume ratio and all that). Memory doesn't inherently have the same problem because you only need to activate a small part of it at a time to get particular bits out. ROMs are particularly nice because they don't need to dissipate power just to remember a bit.

Another reason that big 3D memory is much easier than big 3D processors is because, with memory, areas with manufacturing defects can simply not be used.

Yes I do ...

[satans_advocate]

Anyone else have any good ideas for this chip?

Yes! I've always wanted a ROM image of the entire baseline OS/distro. Kernel, drivers, lib-c, X.

Advantages include....

System files cannot be compromised or corrupted.

Good for thin clients. Put one of these chips on the network card.

Other?

but ..

[dbcad7]

How would you patch it an update it ?

Re:but ..

[zenst]

easy, changes will be stored on disc but you will still need the base rom image.

Worst job ever

[SuperKendall]

Marketing something that can't be described and only seen.

Obligatory DRM / TC inference

[inject_hotmail.com]

Ok, without reading all the -1 and 0 rated posts, I have noticed no one has mentioned one of the most ominous potential implementations of this technology. I am very worried that a technology such as this will make it VERY easy and inexpensive for companies like Blockbuster to mass market inexpensive miniature media encoded with digital rights.

I am worried about this because in effect it will pull the wool over the eyes of the public because joe-sixpack will be overly joyed to not have to return the $3 matrix mediachip movie back to the video store, ignoring the fact that he is ratifying a terrible technology...I mean, what could be better than renting a movie that never has to go back? No more late fees!!! Who cares if it only works for one night? It was only $3!

If they can get the media players to the masses, I'm sure they can require DRM/TC compatibility through a patent or copywrite blah blah blah.

New technology = good
Corporate implementation = not so good

If this does happen, I hope the revolt is just the same as the last one on the matter.

DRM/TC on PC's, different story...it's already here.

Inject
Anti-drm/tc

Exponential?

[Anonymous Coward]

Just wondering, were you thinking when you said that light is exponentially faster?
It is faster, but where have you pulled this exponential crap out of? You can have an exponential relationship between two variables. The speed of light vs the linear speed of a DVD has SFA to do exponents. It makes you sound like you have no clue.
Big number =! exponential.

So then the followup question

[SuperKendall]

How many Matrix red/blue pills fit inside a Volkswagon Beetle?

obligatory ...

[Thoguth]

Guess I'll have to buy the White Album again

Re:one movie?

[Anonymous Coward]

I've seen 8 gig USB memory sticks. Breathtakingly expensive too!

The slashvertizement is for a 8Gb device, the little "b" stands for bit, making it a 1GB read only chip that is not currently in any shipping product.

A 1Gig USB Read-Write stick is available NOW for $70 bucks.

Re:one movie?

[Infinite Entropy]

I MEANT gigabytes. A 'gig' usually is short for gigabyte. Or at least I thought so. http://www.buslink.com/p_cat1.asp?catID=60

Music distribution

[spectrokid]

They don't want me to make backups of my CD's? Well sell me a format which doesn't scratch! Make a small chip I can click into my stereo the way digital cameras work. If it is small enough it should be no problem clicking 20 into a car stereo.

Sweet Bones!

[Synth3t1c]

This will help most anyone out! Cellphones can store more data now, same with thumb drives, palm pilots, and more! There is a lot of potential in this type of memory...

they need better marketers

[sacrilicious]

One 8Gb ROM chip would have sufficient storage capacity to store the contents of an entire movie using H.264 encoding.

Um, seeing as how 2.5+ hours of video compressed with mpeg2 encoding can fit on a *4* Gb dvd, and H264 is a large improvement on mpeg2, it would seem that the above statement is rather modest. I'd bet that eight movies could be stored in H264 on an 8Gb chip.

Re:they need better marketers

[KillShill]

Gb = gigabit

GB = gigabyte

notice the little g and big G.

some people think it's ok to mix the two interchangably...

on board base OS

[displague]

Does anyone remember early PC systems that had memory cards with a read-only operating system? A friend of mine had one that had dos 5 (or the likes) and a sort of literal folder appearance gui with it.

Rant begins. If we had ultra fast, high density ROM chips like this it might be nice to put the core of an OS onto the chip and only use the harddrive (or large RAM) for updated components. A new 'Windows' or 'Linux' system would be inserted into a little cube-tray on your computer . All your 3rd party applications would be left on the hard disk. Hrmm, or, the software could also be purchased on cubes like this. Maybe we end up with a daisy chain of USB2 attached cubes, or a cube-tray, each representing a DVD sized 3rd party application. This sounds more attuned to commercial software. Rant over.

Fragmenting media?

[Grendel Drago]

I don't think 52X was the physical limit of spinning CDs; I think it's the limit of the IDE bus transfer rate, and you get diminishing returns faster than that because of the long spin-up times.

Didn't someone get some mechanical engineering testbench equipment and figure out that CDs didn't start to fragment until 100,000X or something insane like that?

--grendel drago

Re:Fragmenting media?

[Anonymous Coward]

Didn't someone get some mechanical engineering testbench equipment and figure out that CDs didn't start to fragment until 100,000X or something insane like that?

Maybe an unblemished CD

But any tiny imperfection or cracks drastically weakens the CD, making it more prone to catastrophic failure. And a lot of CDs/DVDs develop cracks at the hub area if used in laptop CD/DVD drives (or other drives with the "clicky" mount in the middle).

Some DVD cases are so poorly designed that you're likely to crack the media just removing it from the DVD case.

Re:Fragmenting media?

[Kent Recal]

I know someone who had a CD-ROM scatter in, I think, "only" a 40x drive.
He said nothing really bad happened, just a very wierd, loud noise and the tower shook a bit. So, yes, they actually can fragment at lower speeds - at least when they're damaged as the said disk most likely was.

Design of a chip array.

[Dog135]

Actually, I'd imagine it as a 5.25" drive bay with about 32 little slots (8x4) you can plug these cards into, with each card resembling something like a PQI intelegent stick. Each stick would stick out about .5" so they can be pulled out easily, and the bay would have a flip-up cover to protect the cards and the unused slots. A .5"x.5" label on each card would show what programs are on the card, and be viewable even while inserted into the drive slots.

As far as updating the software on the cards, the format could use blocks with a header containing a "new location" value. If the value is 0 (never written to) then use the current block, otherwise, the data is in the block pointed to by that value. If the card becomes too segmented, have the option to copy the data to a new card.

Even if these card end up being expensive (ie: $5 to $10 each) they may still be worth the money for distributing software on.