Slashdot Mirror
IBM Increases HD Density with "Pixie Dust"
jeffsenter writes "CNET and the AP have stories on IBM's latest major advance of HD density. "Technically called antiferromagnetically-coupled (AFC) media and informally referred to as "pixie dust" at IBM, the innovation introduces a thin layer of the element ruthenium onto the disks inside hard drives where data is stored."" I knew it. Everyone told me through was no pixie dust in computers - but The Truth shall set you free!
Speaking of disk drives and (some sort of) dust: My dad used to work for the only IBM dealership in my hometown back in the 80s for about 7 years. He said in the early 80s there was a lot of people with systems (like those sweet sweet 8088s) that would be scammed by people claiming to do strange regular maintenance. One of my favorite stories was when he said some lady came in and brought in her box and asked him if he could do the regular maintenance of vacuuming up electrons that had fallen off her disks and were filling up her case and making it slower. Apparently this lady had been conned into some expensive cleaning job by someone who pretended to do just such a task and thought that it was a required maintenance. At least the con artist was wise enough to know that disks have something to do with electrons and magnetetic fields. I wonder how much money he made scamming people that way. -tmh
This is true; people are trying to develop, for example, cat converter material that uses only common metals; this usually isn't a problem for catalytic material. The concern I'd have here is that we're looking at pure atomic Ru, and it's very hard to mimic the purely atomic properties by using other elements. Maybe alloys or the like, which have the same amount of atoms, but only a fraction of them being Ru, will have the same effect.
"Pinky, you've left the lens cap of your mind on again." - P&TB
"I can see my house from here!" - ST:
Ruthenium is an extremely rare element, moreso than platinum. It's not valued as expensively as platinum only because it lacks a lot of the useful properties that platinum does have (eg very high melting temperature, chemical resistence), and is usually a byproduct of platinum mining in the first place. I've heard through word of mouth that there is probably no more than ~ 1,000 gallons worth of pure atomic ruthenium known in the world and not currently in use. (Fortunately for us, Ru was considered at one point as a major component in the catalytic converter for cars, but other, more abundent but more expensive materials were found instead.) While each drive that might use Ru in this way would only use a tiny tiny fraction of that 1,000 gallons, a run of drives in the millions could easy make a marked depletion in the supply.
"Pinky, you've left the lens cap of your mind on again." - P&TB
"I can see my house from here!" - ST:
Found a good page on Ruthenium, for those curious about it's uses, who discovered it, etc. 8-)
And it's only $30/g."I'm nobody suspicious... That makes me sound even more suspicious, doesn't it?" - Spike (Cowboy Bebop)
How could this be? That is a quote from the article!!! How can it be offtopic?
Because everyone knows that open source is not magic pixie dust.
--
--
enterfornone - logging in for a change
If Ru is just $30 a gram, and there really is only 1000 gallons of the stuff in the world, it shouldn't be too hard to get all your friends together and buy all of it. Voila! You've cornered the market. Sell it back to them at $100,000 a gram.
So, something doesn't make sense. If Ru was really that rare, wouldn't someone have played that trick on us already?
If tits were wings it'd be flying around.
Given that IBM is a big seller of hard drives with extremely fast interfaces, I'm sure the first drives that use this new coating material technology on the drive platters will be 15,000 RPM drives with LVD Ultra160 SCSI or FibreChannel interfaces and 4-8 MB drive memory buffers. It'll be quite a while before we see this on ATA-100 IDE drives, though.
Raymond in Mountain View, CA
--
The silver-dollar size 6MB disks can hold 6000
minutes of music, 2000 photographs, six hours
of video, and so on.
It's not that programmers are getting sloppy, it's that the designers are getting more ambitious.
:-) ). What takes up all the space is all the graphics, sound and fmv. Now that we have large drives, lots of memory and fast processors and graphics cards, the designers and graphic and sound artists can really let themselves go and create visually and aurally rich games.
The core code of most games probably doesn't take up much more than a few megabytes (I'm at work, so I can't check
The reason that games need more and more disk space is that people like me demand ever larger, prettier and better sounding games. It has nothing to do with programmers not bothering to space-optimise their code. (Indeed, you often have a choice between optimising for speed or space usage; given that choice, I know which will be done)
Cheers,
Tim
It's official. Most of you are morons.
We are eventually going to reach a point where we cannot store any more data per unit area of disk, and we are also going to reach the stage where reading these units of data will be unfeasibly slow using magnetic heads attached to a moving arm. This time is likely to come sooner rather than later, and the problem needs to be addressed.
Well, as you increase storage density on rotating media, your sequential read speed also increases.
What exactly would be the motivation, though? I can see the performance being a big plus, but let's say you bought a 15gig solid state drive.
Then dropped it about 3 feet.
I've seen normal hard disks go wacko over this.
My point being is that mechanical things wear out. I notice over time that the hard disk in my IBM Thinkpad gets louder. There's definitely wear going on, and after a drive dies, we buy another.
Just because a hard disk manufacturer gets 100$ from you for that 40gig drive doesn't mean they wouldn't like another 100$ in 2-3 years.
Karnal
Y'know what makes me happy about this? No mention of patents anywhere.
This is a real advance, they know their competitors can rip their drives apart to find out what they've done and I'm sure they could have found a way to patent it to stop them using that knowledge. But they haven't - they even admit that they expect their competitors to produce drives with this shortly.
Well done, IBM. Patents have their place but you've chosen not to take one out and use it as a weapon - even though it would give you market control. That's good for us all and I thank you.
Greg
(Inside a nuclear plant)
Aaaarrrggh! Run! The canary has mutated!
Actually, that's not being lazy that is programming for lower machines. Uncompressing video on the fly takes a fair number of CPUY cycles, maybe not on your GHz Atlon but on Billy-Bob's PII-300 those cycles saved can let him actually see the video full screen.
It's the same reason CD's contain uncompressed data. At the point of conception, it was cheaper to use the space than to use the processing horse power.
http://kered.org
Well I know I've got a huge handful of old 30 and 72 pin memory simms...Does this sound feasible???
Sure, if you don't mind losing all your data when the power goes out. Me, personally, that's not my cup of cake, though. Besides, your huge handful of simms adds up to what, 256mb? You could get a 256mb ram chip for $50. The price of the IDE bay with simm slots would cost at least that much, and you'd still end up with a bunch of old ram chips in a drive you couldn't upgrade.
I'm guessing not feasible.
What's your damage, Heather?
The sad part of this new technology is that it's going to allow game programmers to be even more sloppy. Back in my day (I'm not that old) I had a 128mb hard drive. Games had to be small.
What?!? So, back in the day, did your game have full-motion video clips? Support for 1600x1200 resolution? Speech clips? Support for force-feedback joysticks and mice? Oh, it didn't? Gee, I wonder why the game sizes are so much bigger these days. Hmmm. Must be because of those sloppy game programmers.
So help me, if somebody mods that post up as insightful, I'll...I'll...
What's your damage, Heather?
make Linux, not Microsoft. sin(beast) = -0.809016994374947424102293417182819
"Today, many PCs come with hard drives ranging from 10 gigabytes to 20 gigabytes." - the AP article.
20 gig is the smallest size of hard drive you can buy! at least, not used! Just on Saturday, I needed to replace a 2 gig on an old server, and I had to go with a 20 gig as that is the only one with an over 1 year warranty. *sigh* old data is so sad...
--
Gonzo Granzeau
Gonzo Granzeau
"Nothing the god of biomechanics wouldn't let you into heaven for.." -Roy Batty
Whilst I don't know exactly what system this uses, my PhD is in magnetic intereactions in systems of reduced dimensionality, so this will work, and wont be far off.
The problem with increasing data density is that neighbouring bits get closer together. There is an interaction between the bits, and the closer they get, the stronger it is.
The antiferromagnetic coupling introduced by the ruthenium (as metalic layer, btw, not as dust), make it more difficult for a bit to be flipped by a neighbour. This allows neigbouring bits to be place closer together, because they can now withstand higher field while maintaining data integrety.
--
No, the first drives with this technology are already on the market. In fact, I have two of them sitting next to eachother in the waterproof, ruggedized,solar-powered Webplayer I'm building. They're IBM Travelstar 20GB 2.5"/9.5mm drives. Nice, quiet, and reasonably fast, but at a mere 4200 RPM they're not what I'd call 'extremely fast'. The magic is used to put 20GB in a small package (there's also a 30GB version, but that one was more than double the price of the 20GB).
--frank[at]unternet.org
Police arrested Edward Peel at 4:30 PM yesterday for grinding up all the harddrives in the data center where he worked. He was found with a metal grinder, and a bucket of ground harddrive chassis, sitting on top of the largest rack. "But I will be able to fly" he told police before being taken away. The remote backup storage firm that employed him has not been available for comment.
-- http://thegirlorthecar.com funny dating game for guys
Of balding Dilbert types in tie-dye suits. Time to up the medication, I guess.
Best Slashdot Co
Oh, about N years.
Best Slashdot Co
Do you need those capabilities for Tetris, Lode Runner, or Arkanoid? :-) There's more to gaming than the flavor-of-the-month FPS.
(FWIW, the game I play most anymore is a Tetris clone called Block Party on my Palm III, and that's only to kill time. I have too much other stuff going on to waste time shooting at computer images; if I want to shoot things, I'll grab my Glock and go to the nearest target range.)
20 January 2017: the End of an Error.
Yeah, it wouldn't be very useful for storage. But I can see how having a big chunck of ram disk space could speed a system up.
/ram, and making sure /ram/bin is first in your $PATH would do the trick. Could do this with main memory too, but it would be nice if we could use old simms and save ourselves that precious ram for running applications.
For instance. How about we take this guys idea, and write an algorithem to check which commands and apps a person runs. Each of them is given a priority, based on how often they are executed. Copy those commands onto the ram disk, and let the user execute them faster from there, without the need for disk i/o. Save the settings in a simple file, and upon log in, the user's frequenty loaded commands are quietly loaded into the ram disk.
I'd be pretty easy to make it work for multi user systems I would think. I guess mounting the ramdisk on like
Okay, so I have 384 megs of ram in this box and I don't need it all, but like hell am I gonna share it!
--Nuintari
slashdot : where an opinion can be wrong.
Sort of, but no trade off on the login time. The restoration at login would be run at a low priority so that it wouldn't be noticeable. Plus, on a multiuser system, stuff that is used by multiple users could be shared, and even, if used by enough users, or some other criteria, it could be left in this space indefinately.
--Nuintari
slashdot : where an opinion can be wrong.
Well..... there's something about taking main system memory and using it for swap.... it seems.... silly. Since swap is virtual ram that the memory management sub system sees as just more ram with a lower priority (not quite, but basically). So, taking ram, and telling the copmputer to pretend its ram.... seems a little silly.
And that's not exactly what I wanna do anyways. Ya close a program, it leaves memory. Open it again, it has to load from the disk. That's slow. So, why not, load it the first time from the disk, and while I am running the program, it copies in the background to this "buffer" level between the disk and ram. So, when I close the app, it leaves memory, but opening it again later results in it being loaded from this weird ram disk array of simms.
I dunno, with today's computers, it might be almost unoticable. Closest I have done to this way using a small ramdisk in a similer way (that required a lot of micromanagement, but it did make stuff faster). Most of my interest in this idea is due to a love of putting old hardware to use, and just to see if it can be done.
--Nuintari
slashdot : where an opinion can be wrong.
I guess that sortof helps. A little.The only clue as to what this stuff does is this tidbit in the AP article:
So, I take it that antiferromagnetically-coupled dust mites increase the blargle factor of a magnetic region? (For some suitable meaning/value of blargle.)
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
I'm not too surprised that they found a new way to increase data density. IBM has a great research department and they sell a lot of hard drives. I am curious to see if my computer can handle a 300 gig hard drive. And it will be cool to see their microdrive benifit from this also.
All I really want to know is WHEN? When do they expect these new drives to hit the market? How much testing do they have to do until they tool up for mass production?
Democrats or Republicans. They are both taking us to the same place and they are not afraid of us anymore.
Ever hear of MRAM? Or how about MRAM? Oh, and don't forget MRAM?
Don't lead me into temptation... I can find it myself.
I'll also buy that this would be really helpful for situations like a notebook where the computer it's used in is frequently shut down in a controlled fashion. Does it buy you any quicker boot in the case of a power failure? My PC running Linux is a complex system in which the processor (registers and pipeline), the cache controller, the cache, the main memory, and the disk (virtual memory) all have to be in a coherent state. Does the non-volatile nature of MRAM really make it possible to recover that complete state, or do I still need to go through something like a reboot? Commercial power failure is far and away the most common cause of a "shutdown" on my desktop system...
I know that Slashdot doesn't employ reporters, like CNN and AP. Nevertheless, in claiming to be "News for Nerds. Stuff that matters", and considering the respect Slashdot has earned as a source of editorialized news filtering, doesn't it seem about time for Slashdot to be a source of direct dissemination of this sort of news concerning technological breakthroughs?
Right now, Slashdot in some ways resembles The Drudge Report. Drudge himself is hardly even a hack in the news space. He is, at best, a right-leaning news filtration system, with a taste for lurid sensationalism (hence his willingness to constantly link to stories which many in the right-wing would call immoral, perverse, and/or otherwise not worth calling news...). He is, at worst, a launderer of political gossip for extreme wings of either political party, most especially the right. (Again, his love of the lurid will often lead him to be used as a tool for spreading gossip for the left-wing, normally against political opponents on the right... But such is the lot of a tool and slave of scandal and gossip.)
Point being: Drudge barely participates in traditional journalism (and thank the heavens for that, considering his absolutely cringe-inducing grandstanding). His existence is truly parasitical. He depends solely on real journalists working for other news organizations, magazines, etc. to create his own brand of "news".
Slashdot differs from Drudge Report in many ways obvious to any reader of this message board. The very fact that there is a "community" component to Slashdot, with peer-review of comments, and further reviews of those reviews, is substantially more engaging to the news reader than Drudge's gossip post. Yet the two news sites share a common trait- namely, that they fashion their own "source" of news that is itself simply a filtered bias towards other news, administered dutifully every day by human decision-makers who understand the values / interests / worldviews of the demographic they are serving.
Now, fast-forward to my point about "Pixie Dust", so I can tie this sprawling mess into one universal point and walk to my fridge and grab myself a Bawls or two...
It seems to me that there would be an incentive for companies like IBM to cultivate a direct relationship with top-tier tech-news outlets like Slashdot, so that they could break news of their own breakthroughs even faster than they currently do. (Just as political operatives have incentives to break news on DrudgeReport, for instant penetration of the radars of the community...) It seems a marginal effort would be required to get releases about such breakthroughs DIRECTLY to slashdot, so that we would be able to see it here FIRST, rather than see it first on CNN, then a couple of other sites, and FINALLY breaking on Slashdot (the "news for nerds, stuff that matters" network) after it is already soon to be pulled from front page rotation on CNN, etc. I'd prefer that my specialized news sources, like Slashdot, were breaking stories about their field of specialty (technogeek news, in this case) rather than rehashing the geek ephemera that generalized news powerhouses like CNN are producing.
I want to see Slashdot evolve so that it doesn't continue to subsist in the purely parasitical manner that Drudge does.
I want to see whatever clout you've created as a news brand parlayed into better access to breaking news, right from the source.
I want to see some evidence that the successes you've had are leading to growth not only in awareness about you but in your power and sophistication as a news site.
Most importantly, after saying all this, please consider my last request as carefully as anything else you've seen in this entreaty, should the fates smile on me and moderate me upward into your scrutiny.
As you grow more and more like a news organization, gaining in power and might, for the love of all that is good to us your readers, please oh please don't put another Jon Katz on your payroll.
That is all.
Let's hope that they don't figure out some way of embedding CPRM into the hard disk surface, too.
Yes, I am joking. I hope...
Ok, so my system DID go down yesterday when the snowstorm caused a power outage. BUT... how about powering it up and loading your entire Linux system on it? Take a minute or two to suck the entire system off the hard drive and load it onto the RAM device, then when you go to start Netscape, Emacs or X (Probably the three biggest offenders on my system) instead of them taking 10 seconds to load, it'd happen instantly.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Get it right, Hemos - The Truth shall make ye fret!
-J
Karma: T-rexcellent.
The monkeys appear to be related to this dust, but the exact nature of the relationship lies just beyond my grasp. Perhaps they have something to do with the genetic experiments conducted by one IBM's former business associates, Germans I believe.
Hopefully these HITS of dust will be available to the mass market soon!
--
--
I like to watch.
This message brought to you by the Carpatho-Rusyn Liberation Front. Accept no substitutes.
--Fesh
--Fesh
Kill -9 'em all, let root@localhost sort 'em out.
High resolution and full-motion video clips do increase dramatically the size of games. However, there are many games that store those clips in uncompressed AVI files. That's being lazy. No matter how big storage capacity is, an efficient programmer must make the best possible use of it, if only to have space for even more video clips.
Today, an entirely solid state computer seems to be still very far in the future. They are even making PCMCIA hard disks for cameras and PDAs, so the trend seems to be moving in the opposite direction.
I've always known they put Magic Blue Smoke into computer components. It's undetectable, but you can definitely prove it's there. When you let the Magic Blue Smoke out of the device, the device doesn't work anymore. This goes for anything that uses electricity. :-)
So everyone wants to give the little hampsters inside our hard drives (or pixies-WHAATEVER...) and many people are whining about "Solid state drives would be Soooooo Coooooolll!"...
Oh...wait...so was I...Well I know I've got a huge handful of old 30 and 72 pin memory simms...No, they aren't as fast as a new DIMM @PC133, but I'll but they're a Helluva lot faster than any old spinning hard-drive...and if I'VE got a couple o handfuls...I can only imagine what some other people have, let alone anyone who has been specifically collecting this stuff....
So what I'm thinking about is like an external drive bay that has NOTHING but SIMM slots in it-and a LOT of em....capable of maybe 20 gigs...okay-maybe 5 gigs...Connect up through a PCI card or even an existing ide cable or something...
Does this sound feasible???
Sig currently under construction. Mind the gap....
Seams to me that Kingston (and probably others) sell flash memory modules up to 512MB. With a little bit of hardware know-how, I think it would be possible to finagle an array of a few of these (let's say, oh, I dunno, 8 should do it) for a grand total of 4.096 GB. Expensive? Yes. At current prices, (mpsuperstore.com says $600 per 256 MB) this is... um... $4800. Yikes!
Anyhow, that's about $1.17 per megabyte. Cnet says I can get a 4.3 gig HD for $91. This is about $.02 per megabyte.
Just from a quick google search, I found out that generically, 512 Mb flash mem supports a r/w speed of 3.0/1.5 Mb/s. I'm going to go out on a limb and suggest that "seek time" would be something like 200 nanosecs. Our 4.3 gigger, on the other hand, maxes out at around 12/6 Mb/s, but only an "average" seek time of 9.5 milisecs. Comparatively speaking, the flash memory wins the seek time test (almost 50 times as fast) while the hard drive is four times as fast reading and writing. Tortoise and the hare, anyone?
The only other thing I can think of is to some how support an array of standard, EDO or SDRAM memory with a keep-alive voltage with a hard disk only for backup. I mean, I certainly wouldn't want to store any important data on volatile memory, but for Quake and the like, you probably couldn't beat it.
Just another random thought: I wonder how fast a computer would be able to boot up using pure solid state? I don't have one of those palm-thingies yet...
Please feel free to correct my math or my (ahem) facts, this was just a quick search during one of my numerous "email catch-up breaks". Thanks.
There is simply too much glass..
Thought I would see the day that "Tinkerbell" was a viable name for a line of hard drives...
Papa Legba come and open the gate
It is too that programmers are getting sloppy. Take Playstation save files, for instance: The original Playstation had a 128kB memory card divided into 8kB blocks, and the majority of games used just 1 block (some used 2), for a save file size of 8-16kB. Now we come to the Playstation 2, with an 8MB memory card, and all of a sudden save file sizes have jumped to 100kB or more. You'll never convince me that's not sloppy programming (or over-design, which is the same thing), at least in the cases I've seen.
Though I won't dispute the argument that game engines don't grow to 60GB... (:
--
BACKNEXTFINISHCANCEL
Don't forget to put a bit bucket under the open cable when you go looking for the token. That way, you can catch any data that falls out, and pour it back in the cable before reattaching it.
I thought I saw a IBM-Employee Clan of Dwarves splat'n Pixies for their dust in greater Faydark! Or did they buy it over on eBay?? along with their BloodForge armor??
LFS. Have you built your system today?
When IBM released their 25Gb per platter drives.
Geek dating!
GPL Deconstructed
Not per platter, but per ^2cm.
Geek dating!
GPL Deconstructed
And if you really want to dive deep, here's the scary level of detail"> you're looking for...
--
--Got Lists? | Top 95 Star Wars Line
The Required posting of theory behind antiferromagnetically-coupled media http://www.aps.org/meet/MAR01/baps/abs/S6820002.ht ml. Interesting Stuff...
--
--Got Lists? | Top 95 Star Wars Line
"Peace, love and Linux"
:)
"Our new hard drives have pixie dust in them!"
Whats next? IBM OpenAcidTab 1.0?
I've long believed that when our network goes down its because the elves have tripped over their shoes laces.
Yo, so a few more related IBM links from this late reporter...
1 8_pixie_dust.shtml
0 01/20010202_mram.shtml
IBM Research's announcement w/ a link to the whitepaper on the new tech:
http://www.research.ibm.com/resources/news/200105
IBM Think Research article on Solid-State RAM (eventually storage, too?):
http://www.research.ibm.com/thinkresearch/pages/2
Si.
Does this mean I'll be able to hold more Bondage Faries stuff on my HD?
For significant applications, single hard drives don't cut it. The seek times involved for ongoing reads and writes make them inefficient. This can be corrected somewhat using non-fragmenting file systems but even so, as I understand, if there's a bunch of little files continuously being modified (like journal files on a DB) the head's going to be jumping around too much and the app is going to be waiting on it. If they're going to pack this much data on a drive, they'd better speed it up.
IBM is great. Lately they seem to always be on the forefront of Hard Drive Technology. Although I do have to agree with an earlier posting about Solid State Hard Drives. Heck, if I had the money I would definately invest in about a dozen of those 3Gig Solid State drives. Of course at something like $5,000 a pop I do not believe that I will be running those in my PC anytime soon.
If you ignore the other uses of a tool, does that make the tool less useful, or you less useful?
The only possible use for a drive this big is making illegal copies of DVDs and CDs. IBM should know better.
I am wondering if IBM is also conducting as much research into solid-state storage solutions as they are into disk-based ones. We are eventually going to reach a point where we cannot store any more data per unit area of disk, and we are also going to reach the stage where reading these units of data will be unfeasibly slow using magnetic heads attached to a moving arm. This time is likely to come sooner rather than later, and the problem needs to be addressed. When one has a 500 TB hardrive, but the disk IO is only 500 MBps, it will start to become problematic, and at the moment the fact that a hardrive requires a spinning disk is also limiting computer design in terms of both data buses and case appearance. A laptop must have a certain amount of space for a hardrive (although it is debatable whether solid-state drives would be any smaller, one must assume they would surely become so), and those motors take a lot more power than a solid-state media is likely to consume. So, while advances like pixie-dust are great in the short term, what is being done in the long term to create a fully CMOS-based computer, where the only mechanical movement is the fans?
A word can paint a thousand pictures