Slashdot Mirror
Buckminsterfullerene Strikes Again - Nanotube RAM
putaro writes "Nanotube based RAM, under development by Nantero, promises to deliver densities of over 1 terabit per cm^2, is non-volatile and faster than current DRAM. The Economist has a nice story. Forget about just kicking DRAM's and FLASH's butt, is this finally the end of magnetic storage as well?"
This guy is right...
"Ford," he said, "you're turning into a penguin. Stop it."
How is that informative? "I hope it would be compatable with existing memory systems, though. It would be nice to just swap out existing RAM for a NanoRAM module and get an instant performance and capacity boost"???
Maybe you didn't notice - it said TERABIT/cm^2. Your current system probably can't handle more than 2Gig of RAM - let alone hundreds and hundreds of gig. Hell, your BIOS may not even be able to handle a HD that large.
I mean, really. 100x faster, and >1000x the storage. Think about that for a minute. Who gives a fuck if you have to toss your HUGE SLOW FUCKING SYSTEM and buy a new one?
If they do pull this off (and I think they're blowing smoke), it will make today's computers look like the vacuum tube machines of yore.
Hope to drop in upgrade... Come on.
Anything that dense and that much faster than existing memory technologies will need a different memory controller. No getting around it.
The fact that AMD built a memory controller into the Opteron is not necessarily something to be happy about. On the one hand, it greatly reduces latency of memory reads/writes, on the other hand you can't upgrade the memory speed beyond what your entire CPU supports -- you have to upgrade your entire CPU. Which means AMD has to redesign the CPU to take advantage of faster (or different types of) memory. And Opterons aren't that cheap yet...
Okay, correct me if I'm wrong, but even though they say it has potential for hold a terabit/sq.cm., I fail to see how it mandates that RAM modules made with this tech hold more than 128/256/512MB like current modules do. The point is that you would have faster, non-volitile RAM that would fit into existing hardware.
:)]
Or, if you prefer, new computers that can still use the older (and likely cheaper, at least initially) silicon memory.
I don't know about you, but I certaintly don't throw out my system every time they make something better. And let's not forget the people who are so anal/paranoid about performance would pay through the nose for a 10% increase in memory speed. I'd also imagine that this non-volitile RAM uses next to no power, making it a great potential drop-in replacement for laptops and other portable devices. A little bit of software tweaking and you could basically have an instant hibernating mode.
Useful for desktops, too... not everyone can afford/think they need a UPS system, but if the power goes out, smarter software would just bring the machine back to where it was with no data loss! Sounds like a perfect nitch for a compatable drop-in replacement to me. What's the total cost to install new RAM and flash the BIOS and/or apply a patch compared to buying all new systems?
<joke>
Of course, Silicon RAM is to NanoRAM as a kumquat is to a watermelon, but that doesn't mean people are suddenly going to stop buying kumquats and switch to watermelons. I'm using this simple analogy an an insulting way of making a point because, from the way you phrased your comment, I can see your sense of humor and practicality is limited. Unfortunately, I've lost track of the point I was trying to make, but I'm sure you'll agree that both watermelons and kumquats have their season, and either or both make a fine addition to any table.
</joke>
[credits to Dr. Science for that last bit
=Smidge=
Well, maybe not a drop in ram replacement, but how about a solid-state ide/scsi drive? Memory blocks, so to speak...
Look, you're missing the point. This would represent a fundamental change in computing. 1 terabit/cm^2. Imagine having 50 GBytes on your wristwatch.
Half a terabyte on your cell phone.
As many terabytes as you can imagine on a laptop that runs for a day because it doesn't have a HD and all the RAM is NVRAM, and it's 100 times faster than your current system.
Really. Think about it. Who gives a shit if you can upgrade your current machine. Did you see the article recently about AppleII users getting together? You'd look as silly as any of them... (no offense - I've run a IIe emulator within the past 3 months, and it was fun; a lot like it will be amusing to be able to store all of silentbozo's files on my cell phone many times over)
Although your right the 1000x thing and 100x thing were potentials they expect within the next few years, right now they said it was 20x faster then the fastest memory on the market today, so it will be bigger capacity although they didn't say how much bigger, and it will be quite a bit faster.
The hardware aside, in software this memory could easily replace the hd and ram, at least on a linux system. The problems I see are that all software is currently designed to use both memory and disk as if they are two separate things... and to use one device in place of both would require a rewrite. In linux talking to hardware is done through the kernel, so changes to the kernel to "emulate" ram could get this hacked into usability fairly quickly. It could allow the amount to use as ram to be passed as a kernel parameter.
"A CPU with a 1nS clock cycle time and a few MB of on chip cache?"
Just to remind you, 1ns == 1GHz. What is the clock speed on the latest Pentium/Athlon?
as I said buried in another thread, in the overall scheme of things this should require relatively small changes, remember we are talking about replacing disk and ram. The kernel handles access to these and can already address very large amounts of memory. The hardware would actually work as a harddrive and would have a swap partition that the kernel uses as "simulated" memory so it doesn't break existing applications that have things like say... variables. Or it could be a file on the disk in which this virtual memory structure exists and you pass the amount of "memory" you want to the kernel on boot... or course MS would fail to address this issue fast enough and finally will be left in the dust.
Interesting, but how do we get rid of them safely without getting the nanotube particles into our lungs etc, or is this not a problem?
A blog I run for the wealth
It doesn't change anything. Like you said you can get this now. I work in security and ANY system that a knowledgeable person can lay hands on they OWN it. As part of my job I do security audits and I have this little stickers that say "I OWN THIS" if I can get close enough to put a sticker on it then the statement is true. I doubt that will every change, but who knows what the future holds.
Encryption: I may not agree with what you say, but I will defend your right to encrypt it...
Actually I think this will be great for extremely secure storage, because of one simple fact, the nanotubs are sensative to light to the point they explode from a camera flash. That means in extremely secure installations with sensative data it's better to lose than have someone else recover they can use this as a failsafe mechanism when physical security has been breeched, you must pass voice and retina authentication, place your hand on the plate and click your heels three times... if your not the right guy then POOF you see a flash of light, data is destroyed.
how are they going to wire the thing? Suppose the nanotubes are grown, properly aligned and so on. How are they going to place the wires between them? AFAIK, the current technology for wiring the chips is exactly the same that puts the transistors, namely the photo-process. Obviously, this si not going to work on the scale of nanotubes.
Replacing SDRAM (or RAMBUS or whatever) with some type of NVRAM will require a whole new approach to security. Otherwise, when you go home at night, what's to stop me from booting your computer (off a CD or floppy if it's reasonably secure), or rebooting it if you left it running but locked, and running an app that allocates a couple gigs of memory without initializing it then lets me browse it?
Technically, you're right. If someone wanted real security, they would have to make some changes to the OS&Apps and/or system acritecture.
But really, you have to look at the reality of your system right now. Unless your disk is encrypted, and you need something like a smartcard to access the data, you're hosed if someone gets physical access to your machine even once.
There's nothing that stops them popping open your case or booting from CD and copying your whole HD onto their Ipod or whatever. Even if the important bit of info they want is your password, they can always install a keylogger, and have that send them an email with your password (or even post is to a messageboard somewhere, just to aviod being traced).
Actually, I'm willing to state a stronger case. You are screwed if someone can get (unsupervised) physical access to your machine. Period. Even if you encrypt everything. There are so many clever things someone could do to your system in order to get your data that you just could never know you're safe.
I mean, even if you have the system wipe passwords from this new RAM on power down, it won't protect you. I could just open up your case, and stop the clock. All of a sudden, none of that stuff designed to wipe your data is working. I can then hook a logic analyzier and pattern generator up to your RAM, and just read out all your data. If your system wipes its RAM too frequently for that, I could just have an ASIC fabbed and put in on a little board which plugs in between your motherboard, and your RAM.
The only way to stop this is to basically turn your RAM into and uber-smartcard, but even then, it's possible to hack a smartcard too.
I guess my point is your thoughts are basically academic. Yes, this tecnology would add another way to exploit physical access to a PC, but there are already so many of those that I really don't think it matters. The only way you're going to get real security from someone with physical access to the system is to encrypt all chip-to-chip interconnections, and use whatever neat packaging technology the military uses for the chips in its military GPS units. Not very likely to happen.
Life is too short to proofread.
Well, if you use an 8 terabyte equivalence for 1 Library of Congress (the actual definition seems to be a bit slippery, and that's the first one I found on a Googling)...
.62 to .73 inches thick, but since we don't know how thick nano-ram is, let's just assume a wafer the size of an iPod).
Theoretical Nano-Ram storage capacity == 1x10^12 bits / cm^2
1x10^12 bits = 1.25x10^11 bytes = 116.4 Gigabytes
That's 0.114 Libraries of Congress per cm^2.
An iPod, according to Apple's website, is 4.1 by 2.4 inches (it's also
1in = 2.540cm
4.1in = 9.840 cm
2.4in = 6.096 cm
Let's chop a cm off each of those to account for the casing, structural bits, and soldering points that aren't actually storage space. That gives us a size of 8.840cmx5.096cm for our hypothetical nanoPod (so on a tangent, how long before some company introduces the new 'e' and starts dubbing products 'nRAM', the 'nPod', 'nTel nSide', etc?). That's a surface area of 45.049cm^2.
Given our previous determination that we can store 0.114 LoC on 1cm^2, we arrive at a figure of 5.136 LoC/i(or LoC/n for nPod, as the case may be).
"If a man hasn't discovered something he will die for, he isn't fit to live" -- MLK, Jr.
First, I have to confess that I am a Materials Engineer and not some ubergeek with a CSE degree.
But it's a definite fact that technological advances are only made possible with the precedence of metallurgical advances.
Silicon wafers today wouldn't exist without the metallurgical backing to create high purity Silicon, Aluminum, and so on.
The point being that with the discovery of the buckey ball, we are entering a new age of history. We're not there, but we're working on it really hard.
Before you toss me out as flamebait consider that each primary age of human civilization is named as a metallurgical Age: Bronze, Iron, Steel. Some might argue that we are in the Silicon Age right now. However, the impact of Silicon is not as ubiquitious as the impact of the discovery of Bronze, Iron, or Steel.
But the Buckey Ball is going to be similar in the scope of impact as Steel or Iron. Why?
- Structural Materials
- Electronics
- Optics
- Aerospace
It's a FUNDAMENTALLY new material product available for the engineers to play with.So far everybody is missing the point including the Economist article. This stuff would replace SRAM. High performance FPGAs from Xilinx and Altera are made of SRAM with refresh in the neighborhood of ten nanoseconds. This would make vast and fast FPGAs possible.
So, instead of merely replacing system RAM or storage this would replace the CPU, the memory controllers, the video card, the sound card --it would be the ultimate SoC platform.
the memory manager and the part of the device manager than manages the hard drive will be one unit!!!
so to get full potential from this we will have to redesign our OSs otherwise if we stick it in the current system it will be a fraction of its potential.
I am the Alpha and the Omega-3
--When components start getting this small, the chances of having an potential error occur go WAY up. What I'd be interested to see, is what they're doing to protect against stuff like cosmic-ray bit pollution and such.
) . That's still 250Gb, or 30 GIGABYTES/cm^2.
... 1/60 of a cm^2. That's .02cm^2, right? Which is about 1.4mm x 1.4mm. And it's fast. And it's NVRAM. I'm willing to take a hit and add some space for redundency.
Why do you say that? Any evidence to back that up?
--After all, if the scale is NANO, one cosmic ray or stray electro/magnetic field can potentially screw up a lot more percentage of memory... Massive redundancy, high speed and constant bit cross-checking would seem to be a reasonable requirement for these chips.
Actually, the article (or related) says that multiple tubes are moved for each bit, which supplies some redundency. But when you get right down to it, let's say you use about 75% of any given chip for parity (Obviously a ridiculous amount - I believe that's a Hamming distance of 11-12 assuming you added the redundancy by the byte (you could correct 30% bit errors in any byte); see also http://www.personal.uni-jena.de/~pfk/mpp/ecc.html
Let's do a little more reverse math. I'll say we want to have 2GB RAM (tops for most of today's desktops). Let's say we don't bother with error correction at all. You'd need a spec of memory
--For just one example, look what a few scratches can do to a CDR - or worse, a DVD. If you can't read it (use it reliably XMillion times) it's basically not very useful...
I seldom take my RAM out of my machine and scratch it.