Scientist Uses Nanodots To Create 4Tb Storage Chip

arcticstoat writes "Solid state disks could soon catch up with mechanical hard drives in terms of cost and capacity, thanks to a new data-packed chip developed by a scientist at the University of North Carolina. Using a uniform array of 10nm nanodots, each of which represents a single bit, Dr. Jay Narayan created a data density of 1 terabit per square centimeter. The end result was a 4cm2 chip that holds 4Tb of data (512GB), but the university says that the nanodots could have a diameter of just 6nm, enabling an even greater data density. The university explains that the nanodots are 'made of single, defect-free crystals, creating magnetic sensors that are integrated directly into a silicon electronic chip.' Dr. Narayan says he expects the technology overtaking traditional solid state disk technology within the next five years."

Comments

How long...

[Pojut]

...until I can get a decent (120GB+) sized SSD that doesn't cost as much as a new video card?

Re:How long...

[davepermen]

who cares? as long as it gives you the biggest perceived boost in performance ever, and thus is worth more than buying a new computer that costs at least 3 - 4 times as much, the ssd still is cheap. which is why i bought ssds for all my systems at home, except the windowshomeserver, as performance doesn't matter there. you can show me any expensive computer out there, and they all are the same to me: slow. believing it to be expensive is not understanding their worth.

Re:How long...

[Anonymous Coward]

But WHY should it not be that price?

SSDs are
1) new (ish)
2) targeted towards portables mostly, or those who require high-bandwidth applications
3) require quite a bit complicated engineering compared to an HDD due to densities

Point 2 is the most important of these, SSDs, while an inevitable goal of computing, are currently targeted to more specialized markets.
These 2 being portable hardware like netbooks, or multimedia / database types who require high bandwidth and high uptime. (not so much with the initial SSDs... oh boy that was a minor catastrophe)
The ones for multimedia and databases usually just went with DRAM drives or RAIDs for speed.
Either way, just like GPUs, they are a very specific market they are targeted at currently, the ones they know will pay more to get a huge boost in performance.

Re:How long...

[Venik]

1) new (ish)

I hope they are newer than the Quantum ESP5000 SSD that has been attached to my Sparc 10 for the past fifteen years.

Re:How long...

[BikeHelmet]

...until I can get a decent (120GB+) sized SSD that doesn't cost as much as a new video card?

2012.

In late 2010/2011 we'll probably see the old Indilinx stock being sold off for cheap. People will jump on it.

Then by 2012 once all the new processes for NAND and controller fabrication have matured, and yields are very high, we'll see controllers like the Sandforce becoming affordable - truly decent performance, if I ever saw any. ;)

I suspect that sometime in 2012 Sandforce SSDs will hit 100GB for $100, or 200GB for $200. I might be wrong, but that's my guess.

I do not think many SSDs will drop into the sub-$80 range. In 2008 we had $70 32GB Patriot WARPs. Now we have $120 30/32GB SSDs - far faster, but more expensive per GB. NAND will drop in price shortly, but not enough to put SSDs in the sub-$50 range. For that price range, we'll have to stick with SDHC/SDXC, and HDDs.

Re:How long...

[swsbai]

Impression births brought in big bad determine along one'study, body of http://www.watches21.com/">replica watcheswork and animation.inward reality,impression dissembles not lone unmatchable heed only in addition to hullos body. Those with life-threatening low level impotent to keep their crop or carry extinct rule bodily function* of imprint. The symptoms mainly let in flavours from sorrowfulness, hopelessness, not wanting nutrient http://www.watches21.com/">replica watchesand ail dormant.Besides, depressive disorder costs linked to more former diseases, care heart disease, What's worse, deprssion is one of the main courses of people's suicide.

Re:How long...

[hitmark]

depends on how fast HP and others can turn memristors into a viable product.

Wow

[MyLongNickName]

My first PC had 4k of RAM. I should be used to this type of growth by now... but it still makes my heart race a bit when I see ever increasing memory capacity in an ever decreasing form size.

I'll tell my grandkids about my first PC and they will roll their eyes as they leave my retirement home...

Re:Wow

[Lumpy]

Problem is most software developers and OS makers also race to consume that memory. Honestly all the software today is a bloated blob that is horribly unoptimized for speed and efficiency.

It's disgusting how bloated most stuff is because we have 4gig of ram and 2 2.5ghz processors... why make it leand and mean? it compiles, ship it.

Re:Wow

[Idiomatick]

It may be peaking soon though. 6nm is getting close to physical maximums for most techniques due to the casimir effect. Techniques that push chips from 2d into 3d will be the next useful improvement. But after that point we have run out of easy options.

Increasing speed of chips and ram could help relieve that pressure mind you. As programmers can tade off more processing for less drive usage, or count on faster ram and compress everything. This will give us a bit more time. Beyond that we will simply have to get more inventive on how we use computers.

Very very fast internet could become important, if users feel they need access to 10million TB of data personally. That may not be physically feasible on a personal computer. So 'cloud' type services would be important. Having a few duplicates rather than 1million duplicates of any given song is clearly a big improvement. This of course feeding into the idea that when we made the internet we stopped making machines, we just started making components for the one ultimate computer. And when you think of it from that perspective there is tons of room for improvement even if some of the parts are nearing the useful maximums.

Re:Wow

[nomorecwrd]

Yeah!... I still remember those crappy games I played on my Atari 2600.

but... with time, graphics, sound, and everything got a LOT better. Same machine, same hardware, same everything.
Programmers finally did incredible things with just 8KB , 1MHz processor
With today's pace of technology, there is no time to think and squeeze the hardware to it's limits, 'cuz once you start getting to know what you have in your hands, it is time to move to the next generation of tech.

Just my two pesos

Re:Wow

[Thanshin]

why make it leand and mean? it compiles, ship it.

And what's the answer to your question?

If it works, why optimize it? To save in storage space? How much would I be saving? 10$ in storage space for every hour of optimization?

It's not art, it's a business. You could as well ask why we don't replace steel by titanium in cars.

Re:Wow

[Anonymous Coward]

There is alot of caching and look-ahead optimization going on in some software that results in alot of data being generated and stored "at the ready". It is all in the interest of improving performance so that you have less lag/stuttering from waiting for network and/or hard drive IO. For computer games there is alot of artwork(models, animations, textures) and audio that is loaded, so of course with higher quality graphics you will see greater memory usage.

Re:Wow

[Anonymous Coward]

in general everything you say is true, esp regarding consumer software.

Some of us are working to shrink some of those codebases in the enterprise space as it pays to shrink your footprint. Why use 5 machines to host 1000 users if you can host 5,000 on 1? The cost savings and scalability are worth the effort. (and yes, I was involved in a project with that scale factor - 1000s of machines reduced to less than 100)

Re:Wow

[lymond01]

I kind of agree, for some things anyway. Microsoft's Office is one of those. Word, Excel, Powerpoint -- they haven't significantly changed since Office 97. I mean, they are what they are. More wizards now, different toolbar, prettier graphs. But Office 97 was enough for 99% of the users. Email is the same way -- why does Thunderbird take 113 MB of memory to run, when it doesn't do much more than the 500K of Pegasus mail back in 1994. Web sites are definitely more complex, but Firefox is running at 350 MB right now...I don't think Half Life 2 consumed that much memory.

There are lots of examples of better programs, but it generally matters how in depth you get with them. Photoshop's layers, introduced in version 3 back in the 90s, made the program infinitely more useful than without layers. Since then...I don't use it a whole lot, but I think I could probably still get by with Version 3: layers, colors, basic filters.

Re:Wow

[Thanshin]

Would you need the 10million TB on yourself at all moments?

Maybe having a fridge sized data storage at home will become standard.

No need for such incredibly high speed communications if it's just for the volume that gets sent from your home computer to "personal" computer (the one you carry).

Re:Wow

[MadKeithV]

Because it takes longer to optimize the application than it takes for technology to overtake you and erase any commercial benefit your optimizations might have.

Re:Wow

[Anonymous Coward]

why make it leand and mean? it compiles, ship it.

Why check for misspellings and comma splices? If it gets past the lameness filter, submit it.

Re:Wow

[Rogerborg]

I know, it physically disgusts me that developers don't limit themselves to writing Space Invaders and Pacman on quad core 4GB machines, but instead chose to actually use all that memory and processing power. Pass me a bucket someone, I'm going to hurl.

Re:Wow

[divisionbyzero]

Problem is most software developers and OS makers also race to consume that memory. Honestly all the software today is a bloated blob that is horribly unoptimized for speed and efficiency.

It's disgusting how bloated most stuff is because we have 4gig of ram and 2 2.5ghz processors... why make it leand and mean? it compiles, ship it.

Sounds like a reasonable outcome of a cost/benefit analysis. Since when is efficiency an end in itself?

Re:Wow

[tuomoks]

It's called betterment, "sometimes, you have to make sacrifices for the betterment of the entire group" - really? Yeah, nobody has been able to show what has been gained since we had for ex. 512KB memory for 2000+ online users? Processing is much faster today, response times 10+ times slower?? Same processing - the business hasn't changed? Nice pictures(?) - actually using nice graphical (and expensive) terminals - about the same! Yes, the price of hardware has gone down and a lot but do we really have to waste it?

Agreed, it really is disgusting to look some(?) code today! On the other hand - I have won all the bets making systems (big / huge ones!) 2 times faster in response times and using half the memory and half the cpu (all three!) last 20 years, easily - LOL! SO - there are some benefits in ignorance, at least for me! Also - a side issue but the bloat has created (is creating!) a huge amount of security problems! Instead of knowing and thinking using (badly) made libraries, objects, APIs, whatever has never been a good idea but I digress!

Re:Wow

[MBGMorden]

Indeed it is business, but it's also marketing. Sure, at home, I like digital photos, watching movies on my computer, etc, but the sad reality is that in your traditional every-day book-keeping style business, we're doing a lot of the same stuff that we were doing on computer 20+ years ago (evidenced by the fact that in a lot of cases the same COBOL programs running the servers back then are STILL running the servers now). It's just that marketing and increasing software bloat have convinced everyone that we now need a 3Ghz Quad Core with 4GB of RAM and a 2TB hard drive to do the exact same tasks that a 75 Mhz Pentium with an 800MB hard drive and 16MB of RAM once did - and with all the bloat it's not even doing it any faster!

The first time I tried BeOS for example was on an 550Mhz Celeron (overclocked), with 256MB of RAM. That system, over 10 years ago, still blows away anything I've used since in raw speed and performance. We should get benefits from newer, faster hardware. Instead we get increasingly lazy programmers and zero net benefit in speed, but with all the negative costs of new equipment purchases.

Re:Wow

[Anonymous Coward]

If it works, why optimize it?

To make it quicker. You may think that you have 1 GB of RAM available and a 2GHz CPU, but the L2 cache in a modern processor will only hold about 4MB--fetching data or insns from main memory might require as much as 100 CPU cycles. And that's before considering virtual memory, which requires millions of cycles to access.

Re:Wow

[Thanshin]

We should get benefits from newer, faster hardware. Instead we get increasingly lazy programmers and zero net benefit in speed, but with all the negative costs of new equipment purchases.

We do get benefits from newer, faster hardware. The possibility of hiring cheaper, less prepared, programmers.

Re:Wow

[Surt]

That's really an urban legend. Most software today is well optimized. It just does much, much more than software did in the past. It uses more memory because many algorithms are trading memory for cpu because memory is cheaper, or memory for disk access because disk accesses didn't keep up with the pace of advancement in cpu and memory (by a couple of orders of magnitude over the last couple of decades).

Re:Wow

[Thanshin]

To make it quicker. You may think that you have 1 GB of RAM available and a 2GHz CPU, but the L2 cache in a modern processor will only hold about 4MB--fetching data or insns from main memory might require as much as 100 CPU cycles. And that's before considering virtual memory, which requires millions of cycles to access.

Making it quicker is not a good enough result to pay the extra expenses of having to hire better programmers.

The current situation (bloated apps that only work on expensive hardware) is the optimal one, in terms of development cost vs results, with the current technology.

Re:Wow

[Anonymous Coward]

My first storage was 80K hard sectored floppies that cost $8/per.

I snarfed up a used 20M HD a few years later for $600 and it lasted several years. Even ended up on a BBS of mine.

First modem was 256pbs but I could only get 128 reliably. But, I could download nice utilities that were 2K. So, avoided giving whole paycheck to ATT (no free LD)

First computer held 4K of RAM.

Do I want to go back? Give me a break.

Re:Wow

[tuomoks]

Yes and no - I agree with Joel if the goal is to benefit (money?) one person or a small group. Now - if the benefits would be for larger group or even nations, whatever - per / use, etc cost structure is totally different. An interesting question, will never be solved, LOL!

On the other hand - I think it also has something to do with laziness and ignorance? Using ready made packages, objects, APIs, etc doesn't require even near the same skills as creating something yourself. Nothing (much) wrong in that, that's the way today, but really, I wouldn't call people doing that "developers" or even worse, "architects" !! They were called at one time coders and when growing up a little, programmers - not much waited from them except following orders, as today!

Re:Wow

[Surt]

You can't do that on today's hardware because of a couple of reasons:

1) the algorithms that are efficient are now well known, and taught at the university level. This means there is less room for improvement.

2) compilers are really good now. So you can't squeeze much out of careful assembly. In fact, there are only a handful of people in the world who can output better machine code than a modern compiler, and only with many hours of investment. So again, there is a lot less room for improvement.

3) the hardware capabilities are too good. With the 2600, you could have only 4 colors on screen. But then people found tricks to get 6 colors, and that was a 50% improvement. Today you can have more colors than the human eye can discern. No trick you can come up with can make any improvement.

Re:Wow

[Anonymous Coward]

Poser.

Get a clue to what he is actually saying. Oh wait you are one of those fake programmers that cant program your ass out of a wet paper bag.

Lumpy is dead on. Write then optimize, not write a prototype and ship it.

Only no skill, ankle biter, wannabe's program the way modern commercial software is written. I have looked at a lot of code over my past 20 years as a programmer. The utter crap you kids churn out is exactly that.

Complete and utter crap. Go back to school and learn from a professor that is not an idiot.

Re:Wow

[Amouth]

You could as well ask why we don't replace steel by titanium in cars.

that is simple - they don't do that or something like that because then your car might last you a long time - and that would cost them money because they wouldn't be able to have you a recurring revenue stream.

Sorry but the accelerated use of plastics and cheap alloys isn't an accident or an improvement in cars..

Re:Wow

[LUH 3418]

Another big improvement may be optical chip interconnects. This could make the connection from the processor to the RAM and other devices much faster, while also saving space on motherboards and RAM chips to put... More RAM. Not to mention possible power savings, and the fact that it should be rather easy to have more RAM channels with this technology... Imagine your processor having an individual, parallel connection to each RAM chip.

It's true that eventually, we will reach a plateau, and in a sense, I think it's fair to say that computers are already evolving less rapidly than they used to, at least in terms of performance... but we're not quite at the limit yet. We might still get to see 50TB SSDs that can do gigabits per seconds, desktops with 512GBs of RAM, processors with 128/256/512 cores, etc. We will get to see personal computers with levels of performance that seem "ridiculously unnecessary" by today's standards.

And once this technology really does reach a plateau, if we really do find ourselves "stuck" at some performance level, it will force people to... You know, optimize things. Come up with clever ways of doing more with what you have. It could also lead to *gasp* more standardization, as the evolution of computing technology slows down.

Re:Wow

[ScentCone]

You used the word "bloated" twice.

Re:Wow

[ProfMobius]

With the current rate of raise in capacity and reduction in size, I think the 10 million TB will hold in a hotel sized fridge or a beer cool box. We will never see full fridge sized home data storage.

Re:Wow

[parlancex]

This attitude angers me. It's a similar evil to pollution or littering; each developer looks at their own software and says it doesn't matter individually if it is wasteful and inefficient, but the totality of all the bloated inefficient software on my computer causes it to take 2 minutes to boot when we're talking about a machine that is technically capable of performing that task in under 10 seconds.

Re:Wow

[Anonymous Coward]

Memory is not really cheaper than CPU once you account for the whole memory hierarchy. The L3 cache only holds a few megs of data, and fetching data from main memory requires hundreds of cycles. So unless you're talking about a loop where data can be prefetched, it's better to pay a CPU penalty in exchange for a smaller memory footprint.

Re:Wow

[StikyPad]

Techniques that push chips from 2d into 3d will be the next useful improvement. But after that point we have run out of easy options.

Just keep adding more dimensions... Duh.

Re:Wow

[tenco]

I pretty much doubt that your P1 could play HD videos, render code with highlighting in a useful speed or display unicode with only 16MB of RAM. Hell, I'm not even sure you could encrypt a harddrive with AES 256bit without greatly reducing it's speed.

Re:Wow

[Anonymous Coward]

Maybe most, but not all.
Here's one: I know someone who wrote and still maintains a commercial app written for a 16 bit dos environment. Runs like stink. But for him and his customers, performance is a requirement.
Mr. uTorrent cares.
Embedded developers care because they have to.

Though in general I lament the state of things as you do.
Software expands like a gas to fill its vessel, contained by the responsiveness/performance found acceptable by the programmer.
As more and more software devs get SSDs, the slower their software will be on machines with HDDs.

Re:Wow

[Michael Kristopeit]

you're wrong for a couple of reasons...

1) the "algorithms" that make individual pieces of software unique are not well known. i doubt they are teaching university students the best way to decide if a credit card transaction is fraudulent.

2) compilers can only work with the code they are given... if you create a process to ship 1 screw to pluto every 100 years, the compiler can't be smart enough to send a shipload of screws and then another when the supply runs low. the compiler can only optimize low level processes... the developer is still responsible for optimizing high level processes.

3) so now you have relatively infinite colors, but the bottleneck is on refresh rate and frames per second... doubling FPS still creates noticeable improvement.

Re:Wow

[Grishnakh]

that is simple - they don't do that or something like that because then your car might last you a long time - and that would cost them money because they wouldn't be able to have you a recurring revenue stream.

What does using titanium instead of steel have to do with cars lasting a long time? As long as you don't let salt corrode them away, steel-bodied cars will last pretty much forever. Here in the southwest, we don't have any problems with corrosion.

Besides, automakers wouldn't bother to apply undercoating if they wanted their customers' cars to rust away.

Sorry but the accelerated use of plastics and cheap alloys isn't an accident or an improvement in cars..

Now this is just plain stupid. Aluminum alloys improve performance in cars greatly by reducing weight, and also by making engines that perform far better. Most plastics are also a giant improvement; again, weight savings.

Re:Wow

[Michael Kristopeit]

It's not art, it's a business. You could as well ask why we don't replace steel by titanium in cars.

if there was as much available steel as titanium, that would make sense... but the problem is that developers build products on top of poorly implemented platforms, and/or don't actively write optimized code from the start. you don't have to spend hours optimizing if it's already optimized.

not sure what you're talking about saving $10 in storage space... if you have a high volume product with millions of users, saving any disk space or execution time is saving millions of people time and money.

when you disrespect your users like that, they don't come back. when you respect them and their time and money, they'll come back in droves. that is business.

Re:Wow

[Amouth]

the titamin/steel was the other persons comment - i was going off of the idea

as for the alloys and plastics - sorry sure they save weight.. but the plastics ALL break down.. they all age poorly.. compared to actual metal parts - and alot of the newer alloys i've run into working on cars - do not last las long..

sure saving weight is important - but you know .. we can save weight in alot of other places than under the hood - when you start adding up the weight increase from the cosmetic parts of cars it is amazing how heavy the interior is.. now weigh that vs replacing simple effective bolts and screws with plastic clips and clamps..

Re:Wow

[Anonymous Coward]

Thing is, not everyone simply uses their computers for email and surfing the net watching useless videos on youtube. I can't deal with the new AIM client simply because I'll be working in Maya, photoshop, mudbox and other 3d / 2d art aplications, but I get an im, it jacks up AIM's cpu ussage to 100% to load up the tabbable chat window.

If programmers made things a little more stream line, like how they were 10 years ago (aim, winamp, windows media player [more like 15 years for wmp]) then I wouldn't need to have a dedicated internet machine and a dedicated art machine in my room

Re:Wow

[Anonymous Coward]

I've taken a fair number of courses. No one really teaches optimization any more...

Re:Wow

[Grishnakh]

as for the alloys and plastics - sorry sure they save weight.. but the plastics ALL break down.. they all age poorly.. compared to actual metal parts

I haven't noticed any breaking down at all in my 15-year-old car.

Yes, most plastics will break down over time with UV exposure, but there are ways to mitigate this: certain additives, keeping them out of the sun, etc. Plastic parts in junkyards, for instance, break down and generally look like crap pretty quickly, because they're out in the sun all day (usually, the windshields and doors are removed or broken). But cars that are maintained properly don't have this problem.

Finally, plastics let you do things that simply weren't possible or easy in the past with metal parts.

And don't forget, they hurt a lot less if your body hits them in a wreck.

and alot of the newer alloys i've run into working on cars - do not last las long..

Like what? I've never seen this myself. Generally, the only "alloy" used in most cars (besides steel, which is itself an alloy) is aluminum, usually for the wheels and the engine block, and these days suspension components. These components don't have any longevity problems I've ever heard of, though the aluminum wheels aren't as rugged, but then again if you're running into curbs or going off-road with a car with aluminum wheels, you're an idiot. For road use, they're great.

when you start adding up the weight increase from the cosmetic parts of cars it is amazing how heavy the interior is..

Sure; that's because there's a lot more interior parts in cars than there used to be.

now weigh that vs replacing simple effective bolts and screws with plastic clips and clamps..

No one these days wants a car with visible fasteners all over the interior, unless maybe it's a stripped-down race car. That may have been OK in the early days of cars, but not any more. The only way to hide fasteners is with retaining clips, which are normally steel, BTW (but used to fasten plastic parts).

Re:Wow

[TheNumberless]

Sorry but the accelerated use of plastics and cheap alloys isn't an accident or an improvement in cars.

There is the benefit that a largely plastic car that deforms on impact absorbs a lot of the energy that would otherwise be transferred to the occupants during a collision. I know I'd much rather be in a squishy modern car than a solid steel behemoth if I'm going to crash into something.

Re:Wow

[Anonymous Coward]

The bad car anology is...

When engineers figure out how to add an additional 100HP to a car at a given level of efficiency, rather than taking advantage by boosting the performance they decide it's now much easier to do things by making the car 1000lbs heavier.

Re:Wow

[wastedlife]

Your problem is that you are expecting AOL to write software that is not broken. Why are you using AIM instead of one of the many free and/or open source alternatives that offer equal or greater features on the same network, in addition to other networks like GTalk, Yahoo, and MSN?

A few examples:
-Pidgin (Windows/*nix using GTK, OSS)
-Adium (OSX using native Cocoa and based on Pidgin, OSS)
-Miranda-IM (Windows, OSS)
-Trillian (Windows/Mac/iPhone, Basic=Freeware/Pro=Payware)

And there are plenty of others.

Re:Wow

[somersault]

This makes me wonder: why are you loading all of your software on boot? Also, you could probably make use of hibernation to speed things up.

My machine does in fact boot up to the login screen in 10 seconds, Ubuntu 10.10 from an SSD :) Another 10 or so seconds (including loading up some extra apps I've installed) to a usable desktop, and that's with only a 1.6Ghz Atom..

Re:Wow

[aminorex]

Market forces dictate software bloat, not some centralized cabal of scheming plotters designing an optimal return on investment. As long as people buy more and more bloated crap-ahem-itunes-ahem-ware, as long as managers get rewarded for their bloat factors, as long as developers get specs incorporating bloat, the trend will continue.

Re:Wow

[cifey]

It depends on how much something is used. Typical case: Hotshot A writes some horrendous code with the database access inside a for loop, works for happy case 1, ship it. Hotshot A takes another job making twice as much as real programmer B will ever see. Programmer B has to fix all the bugs in A's code. In doing so he has to run 1000 tests which take 10 minutes a piece, {QA/customers} all the same. Spend one hour optimizing it and the testing time is dropped to a few seconds and voila, ROI.

Re:Wow

[mswhippingboy]

I hear this complaint all the time - as if today's programmers are sloppier than "back n the day".
Well I say bullhockey.
While it's true that programs today tend to consume much more resources than the days of old, but the reason is not sloppy programming. Back in the day on that system with 4k of RAM, you didn't have a fancy graphical front end, drag-n-drop, fully-immersive, interactive interface complete with audio and video animation. Back then, If you were really clever, you could maybe implement a simple text menu system - there wasn't even a mouse!.
The bloat of today has very little to do with sloppy programming and more to do with all the bells and whistles we expect on any modern commercial application.
If you want to bypass the bloat, you can probably still run your old copy of Wordstar, or maybe even ElectricPencil if you choose, but I suspect the memories of days gone by are a lot more pleasant than the reality.

Re:Wow

[LilGuy]

If everything were optimized the technology business cycle would slow to a crawl and possibly die. If programs were crafted the way they were in the 80s when memory and speed were expensive and scarce we would have an endless supply of space and speed for everything right now and no need to continuously upgrade. I would like to see a little bit of a slowdown in the cycle myself because I've never been able to play the latest and greatest games, but I'm patient enough that I can wait 3 or 4 years to play them with a newly mediocre system.

Re:Wow

[rrohbeck]

If your code requires 30 or 40 GB of memory to run, you'll look at every nook and cranny to save a percent here or there because it directly translates into lower HW cost and/or better performance.

If it's only 1 GB then, yes, who cares.

Oh a desktop, that is. On my netbook I noticed that Chrom{e,ium} is much more efficient with memory than Firefox (and no garbage collection=>no constant page faults) so I've all but abandoned FF.

Next round: ARM based Linux netbooks. Coming soon to a store near you.

Re:Wow

[Idiomatick]

Let me guess, string theorist? You know, adding more dimensions isn't ALWAYS the answer!

Re:Wow

[Surt]

For #1, the opportunities for large gains are almost always at the algorithmic and not the design level. Credit card fraud detection is so far up the coding food chain that you're almost certainly barking up the wrong tree if what you want is performance / memory compactness improvement. If your concern is that fraud detection is too much feature creep, well, that's what I'd say is using up our resources, so we're in agreement there, but my point is that users like features.

For #2, you're incorrect about what compilers are doing these days, at least given your specific example. They specifically are, in fact, looking for opportunities to run code in parallel that is currently written in a sequential fashion.

For #3, the advantage to be gained from doubling the frame rate is relatively small. Most games can run at 60+ fps at good resolutions these days, with things like 8xAA etc all turned on. The marginal improvements in user experience from doubling the frame rate are worthless compared with the development effort required to actually achieve that without sacrificing quality. Again, users like features, and in most cases developers are going to be much better off adding more features before pursuing a comparatively useless bump in the frame rate.

Re:Wow

[Surt]

I meant memory is cheaper than CPU in the algorithmic sense. E.g. trade a GB of memory to convert an O(n^2) algorithm to O(n ln n). A make your computation take seconds instead of days type of trade-off.

Re:Wow

[StikyPad]

HERETIC!

Re:Wow

[Anonymous Coward]

for #1, you're incorrect. credit card fraud detection is COMPLETELY about performance. it has to execute in real time, or the fraud is left unchecked.

for #2, you're incorrect. like i pointed out, processes that happen relative to a data source outside of the control of the compilers memory management can not be optimized by the compiler. if you write an algorithm to query a database 1 item at a time instead of querying for all items at once and looping through the result set, the compiler can not optimize that.

for #3 you're incorrect. there is much to be gained from increased realism in lighting and texture effects relative to frames per second.

Re:Wow

[jmorris42]

> If it works, why optimize it?

Well just to give one example....

Had the Free Software folks put more than lip service into the notion that Linux/* was a good way to keep older hardware in service, i.e. rejected the list below, when the EeePC shipped with Linux it's specs would have been so lowball Microsoft would have had to drug Win98 out of storage to offer a competing product that would run on em. We would have a firm grip on the bottom of the ladder and be working our way up to Total World Domination.

Bad ideas just from looking as "ps ax" on the F12 install I'm typing this on:

1. 2.2MB resident set for modem-manager.... when I don't have any sort of modem installed.

2. 3.7MB resident set for CUPS to listen for network printers and print jobs.

3. 616KB resident set to process ACPI events

4. 1.4MB resident set for wpa_supplicant when I'm plugged into ethernet and don't even have wifi associated to anything.

5. And don't forget 5.4MB for NetworkManager to manage a wired ethernet connection..... totally seperate from the 1.4MB the actual dhcp client is resident in.

6. 4MB resident set because some process might emit an email and sendmail needs to be ready..

7. Toss in another 12MB resident for policykit, devkit and dbus.

8. 7.5MB resident for pulseaudio.... with no sound active.

Believe it or not, I actually used a 486SX-33 notebook with 40MB of ram in the long long ago. It ran RedHat 4.x and Netscape. A bit slowly but it was usable once everything settled down. The list of plumbing bits above wouldn't fit in 40MB and thus begin furiously swapping before even trying to load X so we can forget trying to run a browser.

Re:Wow

[hesaigo999ca]

I tend to agree as I see bad programming all the time and bloating of all sorts, but really not surprising that M$ pushes for that sort of thing, as the more demanding the software the more you buy hardware to keep up, most of the time also forcing a new OS and the cycle continues...really a bummer when I can still get my tiny p2 laptop to do all i need it to do....and they sell about 50$ each now...

Re:Wow

[Surt]

#1 is just plain wrong, and misses my point entirely to boot.

#2 is true. There are certainly things you can do that the compiler can not beat into submission. But the examples given so far do not reflect anything close to what I would have described as 'bloat', and in any case rarely go unaddressed in most end-user software from any medium sized & up company.

#3 is not true. It's a false trade-off. Those features are not dependent in performance.

Re:Wow

[Anonymous Coward]

I wouldn't mind the bloat so much, if only the software WORKED CORRECTLY. It's not so much that they don't spend time to optimize it - they hardly spend time to TRY it, let alone test it properly.

Re:Wow

[MadKeithV]

On the other hand - I think it also has something to do with laziness and ignorance?

Laziness is the beginning of efficiency.
Good optimization is done ONLY on the current bottleneck, because you want to be too lazy to address the other parts until they really are the problem. Ignorance, maybe. I have noticed that a number of concepts (pointers, and with it memory management are the main ones right now) are slipping out of common developer knowledge.

Using ready made packages, objects, APIs, etc doesn't require even near the same skills as creating something yourself.

No, and it might just free up time to work on problems that have *not* been solved yet. Reinventing the wheel is useful to some extent, to learn how the whole thing works. But keep doing that for years and years, and you'll have the perfect wheel while everyone else is flying airplanes.

Re:Wow

[zwarte piet]

In fact, alloy wheels are often made of magnesium.

Re:Wow

[Grishnakh]

I wouldn't say "often". There's been some BMWs and the like with magnesium wheels, but the vast, vast majority of them are aluminum.

Re:Wow

[geekoid]

"As long as people buy more and more bloated crap-ahem-itunes-ahem-war"

How much did you pay for iTunes? I got mine for free. Also, what are you comparing it to? I am unaware of anything with all the same features that sin't similar in size and response.

ROI is a huge factor in Bloat. Not a lot of companies willing to pay for another 6 months of refactoring to optimize code. And market forces don't enter in to it because the difference in terms of speed is almost never noticeably anymore.

Anyplace where it actually has a real world noticeable impact still optimizes it's code.

Re:Wow

[serialband]

We do get benefits from newer, faster hardware. The possibility of hiring cheaper, less prepared, programmers.

Race to the bottom; here we come.

4Tb of data (512GB)

[Anonymous Coward]

Is that a new standard being used by the hard drive industry? Now, you can have FOUR Terrabyfes (very small font: 512GB) for $99.99

Re:4Tb of data (512GB)

[biryokumaru]

4 Terabits = 512 Gigabytes.

Somewhat misleading? Yes. Inherently false? No.

Re:4Tb of data (512GB)

[Anonymous Coward]

Nah. If it were the hard drive industry, 4 Tb would be 500 GB.

Re:4Tb of data (512GB)

[NotBornYesterday]

b (small "b") == "bits"
B (big "B") == "Bytes"

The hard drive industry isn't out to screw you. AFAIK, HDD storage has always been quoted in base 10 instead of base 2 (K=1,000 instead of 1,024, etc), but the difference was never really obvious until lately as the numbers got huge.

Re:4Tb of data (512GB)

[darthflo]

Talking about the capacities of single memory/storage chips, using losercase b (bit) figures has been the standard for years. Since only techies who care about the capacity of the actual chips read this, it's not that much of an issue.
As soon as you're talking about an assembled product (be it a RAM module, SSD or even a smartphone), it'll be B (for bytes) again.

Re:4Tb of data (512GB)

[cyfer2000]

Game industry has been using bit instead of byte for years. And for memory chips (not memory modules), it has always been bit.

Re:4Tb of data (512GB)

[V!NCENT]

Hard drive industry? This is research... Given that each crystal is a bit, one terabits (one tera crystals) can be packed onto 1cm.

What's the problem?

Re:4Tb of data (512GB)

[Anonymous Coward]

Or even tebibits and gibibytes, if you want. I suspect you don't, but they really make things clearer at times.

Re:4Tb of data (512GB)

[Anonymous Coward]

I like the french word for Bytes. Octet. So there is no confusion between 4Tb and 4 To.

It's probably too late to change bytes to another word in english ;)

Re:4Tb of data (512GB)

[thijsh]

Nah, 4 Tb would be 454.7 Gb.

Re:4Tb of data (512GB)

[Culture20]

4 Terabits = 512 Gigabytes.

How much is that in Terribibits or Gibibybytes?

Re:4Tb of data (512GB)

[Anonymous Coward]

I think you made an error somewhere. I got this:
4Tb = 465.66 GB (or 500 marketing GB)

Re:4Tb of data (512GB)

[camperdave]

Talking about the capacities of single memory/storage chips, using losercase b (bit) figures has been the standard for years.

Yes, and all it takes is one uninformed journalist, editor, or marketing agent to turn b into B, just like they turned 65536 into 65K back in the 1980s. You can't trust anything like this on the web unless it comes directly from the manufacturer, or they explicitly write out the words bits and bytes.

Re:4Tb of data (512GB)

[Anonymous Coward]

4 Terabits = 512 Gigabytes.

You're thinking of tebibits and gibibytes. The ratio is 4 to 500 with the units you used.

http://en.wikipedia.org/wiki/Tebibit

Re:4Tb of data (512GB)

[steelfood]

This is, in fact, the calculation used in the summary.

Bits are delimited in powers of 10, whereas bytes are delimited in powers of 2, hence:

4 Tb = 500 billion bytes.

In reality, 4Tb is more like 466 GB.

But practically speaking, with error correction and other overhead, 4Tb is closer to 400GB of useable space.

Re:4Tb of data (512GB)

[ProfMobius]

Took me a looonnggg time to understand the english notation (I'm french). It is way easier to have two distinct names for 1 bit and 8 bits (an octet).

For example, for this article, I really thought they did manage to put 4 Go per chip before reading the comments ;)

Re:4Tb of data (512GB)

[Dragonslicer]

I like the french word for Bytes. Octet. So there is no confusion between 4Tb and 4 To.

It's probably too late to change bytes to another word in english ;)

Either the French changed recently, or it didn't have a word for byte until recently. A byte is not strictly defined as 8 bits, it's just that the dominant architectures used 8-bit bytes. Other older architectures used other byte sizes.

Re:4Tb of data (512GB)

[wastedlife]

Octet has the same meaning in english, it means a group of 8 things. However, the word byte does not mean 8 bits(though that is by far the most common), it means a sequence of bits processed as one unit of information. It can also mean a unit of information in a computer that stands for a letter, number, or symbol. There are some architectures that use a 10-bit byte, including measurements for SATA bandwidth. Are those called "octets" in france? That might cause even more confusion.

Re:4Tb of data (512GB)

[vegiVamp]

I don't know what you'd consider "recent", but I've known about "octet" for 10+ years. The french have the annoying habit of translating words, which makes it exceedingly annoying to read their code and comments.

I had a cow when I first saw variables named $clef, $valeur and $chaine.

Re:4Tb of data (512GB)

[Dragonslicer]

"Recent" would probably be the 1970's, maybe early 1980's. Once IBM stopped having legitimate competitors, the 8-bit byte became the de facto definiton.

Re:4Tb of data (512GB)

[vegiVamp]

Well, the word apparently is also valid english, and wikipedia puts the earliest known occurrence in 1974, in RFC 635.

Re:4Tb of data (512GB)

[Dragonslicer]

Right, "octet" is perfectly fine for an 8-bit byte. But if the French word for (8-bit) byte is octet, what would be the word for a 6-bit byte? "Octet" is a subset of "byte"; an octet is a byte, but a byte is not necessarily an octet.

Re:4Tb of data (512GB)

[vegiVamp]

I know, and I agree. Still, that's the word they use :-)

I'm not aware of any other word being used, but then again I don't know any french oldtimers or specialists that might have worked with non-8-bit bytes.

Re:4Tb of data (512GB)

[duh P3rf3ss3r]

A nanodot is 1/1000 of a microdot. I suppose that will work for memory. All I know is I did a microdot a few times and had trouble remembering much of anything afterwards.

Re:4Tb of data (512GB)

[Massacrifice]

If you're gonna go with 6 bit bytes, might as well invent new words to go with it.

Re:4Tb of data (512GB)

[Dragonslicer]

If you're gonna go with 6 bit bytes, might as well invent new words to go with it.

I would assume that "sextet" would be the specific term.

Re:4Tb of data (512GB)

[wastedlife]

Non-8-bit bytes are not just a thing of the past. SATA bandwidth is measured in 10-bit bytes. Granted, this is due to the encoding scheme, which contains synchronization information so that a separate clock signal is not required.

Re:4Tb of data (512GB)

[vegiVamp]

I didn't know about the SATA thing, but that *is* why I said "or specialists" :-)

Re:4Tb of data (512GB)

[wastedlife]

Thanks, I had a logic parser error and missed that somehow.

4tb != 512gb

[dch]

It appears that the article does not even say 4tb, just that the device can hold 512gb

Re:4tb != 512gb

[Abstrackt]

It appears that the article does not even say 4tb, just that the device can hold 512gb

4Tb == 512GB. Terabits, not terabytes.

Why the hell they would measure in Tb instead of GB is beyond me though.

Re:4tb != 512gb

[Anonymous Coward]

That's 4Tb (terabits, not terabytes), or in other words 512 GB.

Re:4tb != 512gb

[Frenchman113]

4 Tb is 512 GB.

Re:4tb != 512gb

[PalmKiller]

I think this is where they lost you b=bits, B=bytes...it takes 8 bits to make a byte They did not say 4TB, they said 4Tb (4Tbits) = 512GBytes

Re:4tb != 512gb

[Anonymous Coward]

Because memory is always measured in bits? Not for the modules you buy for your PC, but if you ever bought it by the chip, you bet your ass.

Re:4tb != 512gb

[jimbolauski]

I know a nybble is a much better unit of measure.

Re:4tb != 512gb

[Abstrackt]

Because memory is always measured in bits? Not for the modules you buy for your PC, but if you ever bought it by the chip, you bet your ass.

Well, I just learned something new. Thanks. :)

Re:4tb != 512gb

[John Hasler]

> Why the hell they would measure in Tb instead of GB is beyond me though.

Because each dot stores one bit. They are building chips with arrays of dots, not complete hard drives.

Re:4tb != 512gb

[V!NCENT]

Given that one crystal represents one bit, it tells you (as this is research, you know...) that there can be 1.099.511.627.776 crystals per cm (density, much?).

Re:4tb != 512gb

[shoehornjob]

That's funny. I just assumed that the 512Gb was the L1 cache. It's not that unreasonable to have 512Gb of cache on a 4Tb drive right?? Considering cache memory is usually more expensive than the same amount of ram I'm guessing that we're out of luck. Bummer!

Re:4tb != 512gb

[bennomatic]

thAt i5 th3 pR0blm wltH t3h kIdZ Th3z dAYz. grAmmAr, PuNCtUaTioN and C4pitalizaSHUN r all ARBITRARY.

Re:4tb != 512gb

[caseih]

No, it doesn't matter whether you use SI or the silly gibi, mibi prefixes. 4 Tbit is not 512 GB or 512 GiB. It's 500 GB or about 476 GiB by my calculations. 4000000000 / 8 is 500000000 bytes.

Re:4tb != 512gb

[roachdabug]

thAt i5 th3 pR0blm wltH t3h kIdZ Th3z dAYz. grAmmAr, PuNCtUaTioN and C4pitalizaSHUN r all ARBITRARY.

u 4got spelling lol!!

oblig.

[Anonymous Coward]

http://xkcd.com/678/

The rest is just business, which is easy right?

Not a "chip", merely a "chip".

[Anonymous Coward]

They have microdots at a 4Tb-per-sq-inch storage density; they don't have any controller that can read and write it.

This has been "accomplished" numerous times with holographic storage media before. They just never made the read-writers...

Re:Not a "chip", merely a "chip".

[clone53421]

Correct.

“The next step is to develop magnetic packaging that will enable users to take advantage of the chips,” says the university, “using something, such as laser technology, that can effectively interact with the nanodots.”

They have a storage medium with nothing to read or write it... yet.

Although they seem confident that this will come with time, it’s a bit early to be celebrating. Interesting technology, but time will tell whether it’ll ever be usable.

Re:Not a "chip", merely a "chip".

[0123456]

They have a storage medium with nothing to read or write it...

The perfect DRM! They'll make billions!

Re:Not a "chip", merely a "chip".

[John Hasler]

> They have a storage medium with nothing to read or write it... yet.

Put the dots on a "disk" in rings. Call them "tracks". Spin the "disk" and access the dots by scanning a laser radially so that it can read and write the dots in each "track" sequentially. There just might be some existing technology that could be adapted for this...

Re:Not a "chip", merely a "chip".

[Real1tyCzech]

That would suck. Spindle drives are already too slow. Let's use something a tad faster...please?

Re:Not a "chip", merely a "chip".

[MadKeithV]

The perfect DRM! They'll make billions!

Yes, billions of small nanodots. Until their budget runs out.

Re:Not a "chip", merely a "chip".

[camperdave]

That's idiotic. A pair of micromirrors will be able to point a laser at any point on the chip with far smaller seek times than moving the entire chip. Besides, CDs and DVDs are recorded in a spiral, not rings.

Re:Not a "chip", merely a "chip".

[JustinOpinion]

Actually despite the fact that the summary and article talk about this as though it is an SSD technology, I think it is more likely to be implemented in a conventional spinning-disk hard drive first.

As I recently commented, the hard-drive industry is having a hard time shrinking the magnetic domains on conventional hard drive platters, which use a magnetic thin film. (You can make domains smaller, but they start interacting with one another and not maintaining their magnetization properly.) One proposed solution is to actually pattern disks with individual magnetic dots. The separated dots would interact less and should maintain magnetic orientation better.

There are challenges involved in tracking such small dots in a conventional hard drive, but those challenges are being addressed. In the short term it would be quite a bit easier to integrate a new magnetic-nano-dot technology with existing spinning-disc technology, rather than trying to invent a whole new read/write system. This would also help the nanodot technology be refined/matured. According to this press release, this new technique is a pulsed-laser-deposition technique that is compatible with conventional silicon wafers. In other words, it is compatible with existing chip-making processes, and so we can imagine a later stage where the magnetic dots are memory elements integrated directly into microchips (thus, SSD).

Re:Not a "chip", merely a "chip".

[John Hasler]

> Let'sLet's use something a tad faster...please?

They'll put a transisitor over each dot and couple it to the dot in some way so that it can be read and written. Then they'll add a matrix of metallization and logic to multiplex access to the transistors. Add decoding logic and drivers and you've got nonvolatile RAM. And your bit density has gone down by an order of magnitude or so. Still very useful, though, if it's fast enough. Nonvolatile RAM with densities and speeds similar to those of DRAM would be a real breakthrough. Add costs per GB similar to those of rotating media and you've got something that will fundamentally change computing.

Re:Not a "chip", merely a "chip".

[xaxa]

Magnetic discs (hard and floppy) are recorded in rings (aka cylinders, IIRC).

Re:Not a "chip", merely a "chip".

[John Hasler]

> That's idiotic.

No it isn't. It's simple, robust, leverages existing technology, and is capable of transfer rates of 1000 Gb/sec.

> A pair of micromirrors will be able to point a laser at any point on the
> chip with far smaller seek times than moving the entire chip.

I guess that's why CDs, DVDs, and BluRays aren't spun.

Yours is an interesting approach, but there may be a reason why it has not been implemented for any of the existing optical technologies. The latency would be better than that of spinning media but still far too great for RAM and the improvement might not justify the complexity. After all, the block you need is almost always already in RAM in the buffers anyway.

> Besides, CDs and DVDs are recorded in a spiral, not rings.

Did I mention CDs or DVDs? Think hard disk with a laser instead of a magnetic head. Did you notice that these are magnetic dots?

Re:Not a "chip", merely a "chip".

[Real1tyCzech]

Right now, anything with a decent $/GB that is on par with the better SSDs currently out will change things up pretty well.

I have 2 Dell D6400 laptops here, both with Win7Pro, with identical CPU/RAM/GPU. One has an SSD Drive, the other, a spindle drive (7200 RPM).

Without a doubt, the SSD Drive boots faster, opens apps faster, and reboots faster. "Faster" is actually a poor word to describe it. My jaw hit the floor.

With a few *minor* tweaks, we got that baby to boot, complete to an interactive desktop, in 17 seconds. (~10, not counting POST..which could be *much* improved with a little EFI love...) This is vs. ~30 on the other system, post POST.

Cannot wait for prices to drop any more. Starting to upgrade my personal systems in the next few months... (damn my having to see it in action...I was blissfully unaware and happily saving money...)

I hate this...

[LWATCDR]

This sounds really cool but the artical that it links too is really short on details.
Things like speed? Storage life? How many read write cycles before it wears out? Addressing? is it byte level or page level?
I mean is this only a replacement for flash for is it a replacement for ram?
Cool but it just ticks me off. It is just a tease.
Yes they may not have those answers but I would be nice to know what they don't have answers for yet!

Re:I hate this...

[osgeek]

They don't have any of that information because they don't know any of it. They only have a way IN THE LAB to put a shitload of nanodots onto a medium. They mentioned that they have no packaging (way to read or even really write data into the dots) for an actual product.

It's like Ben Franklin saying, "Okay, I've discovered electricity. Computers should be along in about five years."

Okay, it's not that bad, but I hate that five year timeline that is rarely questioned but is thrown out to lure in investors and grant money.

Slashdot should have an automatic filter that looks for the five year estimate and flags with some "fat chance" special color.

Re:I hate this...

[LWATCDR]

That is why I hate this.
It reminds me of those Popular Science stores.
Or even better the one that sticks in my mind. The THOR drive from Radio Shack.
http://en.wikipedia.org/wiki/Thor-CD

I was so hyped by this in 1988 it sound so cool and it was only a few years away...
It never came.
On the bright side we did eventually get CD-Rs and even CD-RWs but not for a good long time after the THOR drive was announced.

Re:I hate this...

[osgeek]

Yeah, that Thor drive was some great vapor. My painful promise memory, was hologram storage. Back in 1992, I remember holding on to a hologram storage article from MacWeek that described what was supposed to be a consumer product in a year or so.

The media was the size of a credit card and they promised it would hold 100x as much as the current best hard drives of the day. It's a real shame because you just know that there was some fairly fraudulent monkey business going on to motivate guys like that to hawk something they had no chance of ever delivering, much less in the time frames they touted.

It's amazing how the same fraud is passed along by Slashdot every six months in the form of a new holographic storage device that's going to revolutionize everything. It's probably the same core of fraudsters forming new companies and recycling their same tired story to pull in new investment suckers.

Re:I hate this...

[LWATCDR]

Oh yea I remember that as well.
Good times.

Re:I hate this...

[Anonymous Coward]

actually your analogy is just too far fetch. we live in a world that moves faster than that. get with the times, because they just blew right by you old man.

And thus it begins

[Anonymous Coward]

Cue the "that's almost big enough for my entire porn collection" comments in 3... 2... 1.

Re:And thus it begins

[Idiomatick]

This is /. 512GB doesn't come close.

Re:And thus it begins

[the_fat_kid]

512GB big enough for an entire porn collection?
You must be new here...

Re:And thus it begins

[Bucc5062]

I think some folks at the SEC may be interested in this technology. Less obvious then boxes of CD/DVD labeled "Hot babes" lying around the office.

Re:And thus it begins

[Antity-H]

would a beowulf cluster of these be enough ?

Re:And thus it begins

[the_fat_kid]

oh, touché.

read write speeds? # of r/w times?

[peter303]

Some promising capacity technologies could not match standard magnetic technologies in these aspects. On the other hand, early CD ROMS and flash was promoted as "write once" technology. They would be so large large that you did not need to reuse storage.

Oh, great...

[InfinityWpi]

... means I'll have to buy 'The White Album' again...

Re:Oh, great...

[shoehornjob]

WTF. The White Album....where did that come from. Reference please.

Re:Oh, great...

[TPJ-Basin]

Men in Black. See it.

Re:Oh, great...

[gmhowell]

It was a mistake to buy it the first time. Talk about self-indulgent. Ugh. However, I'll happily buy DSoTM and Abbey Road again.

Re:Oh, great...

[geekoid]

the quote or 'the white album'?

The White album was an over hyped Beatles Album.

The quote is from Men in Black.

NCSU != UNC

[fred_sanford]

BEGIN RANT Seriously, North Carolina State University (NCSU in Raleigh) is not the University of North Carolina (typically in reference to Chapel Hill). One is a school (that I happened to have attended twice) that focuses primarily on Engineering and Agriculture and the other is a liberal arts school down the road. Seriously, fact-check much? http://www.mse.ncsu.edu/CAMSS/bio1.html END RANT

Re:NCSU != UNC

[gotfork]

BEGIN RANT Seriously, North Carolina State University (NCSU in Raleigh) is not the University of North Carolina (typically in reference to Chapel Hill). One is a school (that I happened to have attended twice) that focuses primarily on Engineering and Agriculture and the other is a liberal arts school down the road. Seriously, fact-check much? http://www.mse.ncsu.edu/CAMSS/bio1.html END RANT

UNC now also has a cleanroom and the physics department does a lot of nanotechnology, but this guy is definitely at NCSU.

Re:NCSU != UNC

[Anonymous Coward]

Let's just call them both 'Duke University' and be done with it.

Re:NCSU != UNC

[Anonymous Coward]

No, Duke == SUNY Durham.

Re:NCSU != UNC

[Anonymous Coward]

And the frick cares? I swear, you people will "rant" about the most inane crap. Seriously, STFU.

Re:NCSU != UNC

[Crimey McBiggles]

Someone modded this two to save on embarrassment. Seriously, fix the headliner.

Re:Nanodots? ..b-bots?

[GungaDan]

These nanites are evolving at a tremendous pace!

tiny pages

[flyingfsck]

The prof needs a basic lesson in math.

A pain text page of ASCII text is typically about 2000 Bytes, so his 4Tb will store only about 31 Million pages. ...and don't use MS Word for that.

Re:tiny pages

[MyLongNickName]

A pain text page of ASCII text is typically about 2000 Bytes, so his 4Tb will store only about 31 Million pages. ...and don't use MS Word for that

I am not sure what a pain text page is. But I believe you don't know the difference between Tb and TB.

Disks?

[pitchpipe]

Solid state disks could soon catch up with mechanical hard drives[...]Dr Narayan says he expects the technology overtaking traditional solid state disk technology within the next five years.

Is shape important in Solid State? It almost seems as if the article is confusing Hard Disk Drives with Solid State Drives.

Re:Disks?

[Anonymous Coward]

To be honest, "disk" has become a generic term for "mass storage media" whether it's round or not (and "drive" for "device" whether it has moving parts or not).

What bothered me more about the article was that he suggests maybe using lasers to read and write the bits. Troubling for multiple reasons. First, there won't be a 1:1 ratio of lasers to bits (even if it wasn't expensive, the laser isn't going to physically be 6nm, so there goes the great density). So that means moving parts, which means it'd probably have to be more like a disk with read/write head than a chip. And lasers aren't cheap, so that probably means just one laser, which means just one rw head, which means performance more along the lines of optical drives than magnetic drives, which means nowhere close to flash. And for a single sided spinning disk in a 2.5" drive that won't be out until at least five years from now, it might not even be much higher capacity than flash will be at then.

Second, um, using a laser to flip 6nm bits means a 6nm wavelength, doesn't it? X-ray lasers. Not exactly mass production stuff yet.

Re:Disks?

[Ant P.]

The third letter in "HDD" is only there because it has a motor. SSDs should really be called Solid State Storage, or SSS for short.

Re:Disks?

[Anonymous Coward]

Drives? No we haven't used hard drives in over 30 years. We store everything on 2.5" disks that can be rolled up like paper. Not sure where you have been but I think you need to report in for government mandated updates and "training"

3 ... 2 ... 1 ...

[NotBornYesterday]

I suppose that depends on which video cards and SSDs you use.

Re:3 ... 2 ... 1 ...

[tenco]

That's no video card, that's electric heating. Would be interesting if this gtx 470 gives one more computing power than a (beowulf ;) cluster of dual core atoms for the same TDP. Transistor count is close enough.

Re:3 ... 2 ... 1 ...

[Hurricane78]

The problem with SSDs is:
1. cheap
2. big
3. reliable
Choose two!
But even then, you can only be sure of number 3, after some years have passed. For obvious reasons of there not being any test data for years of use, until years of use have passed. ^^

Re:3 ... 2 ... 1 ...

[Kjella]

1. cheap
2. big
3. reliable
Choose two!

Please tell me where I can get cheap big SSDs.

The advantages I've seen:
1. Performance
2. Silent
3. Light
4. Higher shock resistance

Undecided/even:
1. More reliable
2. Smaller
3. Lower power

Disadvantages:
1. Cost
2. Size

Unfortunately, one of the things often touted about SSDs don't quite seem true, they don't really scale down. Big SSDs = more parallel channels = better read/write performance. Which is a shame, because if they could really scale down to 10-20 GB I'd see a huge market for dual storage laptops, SSD for OS and HDD for bulk storage.

Re:3 ... 2 ... 1 ...

[cowboy76Spain]

But even then, you can only be sure of number 3, after some years have passed. For obvious reasons of there not being any test data for years of use, until years of use have passed. ^^

BEGIN PEDANTIC

In fact, if you assume a Poison distribution of failure (where two identical working chips are equally probably to fail, no matter that one of them is brand new and the other has had years of use), you could just put 10.000 to test, and, if after a year 100 of them did fail, rightfully claim that they have a MTBF of 100 years.

Of course this distribution does not fit other components (mainly those with mechanical parts -CDs, HDs-) because there is progressive wear so an older unit is more likely to fail than a new one, but maybe it works for pure chips (anyone has info on the contrary?)

END PEDANTIC

Re:3 ... 2 ... 1 ...

[Beezlebub33]

Poisson, please! (and in a post labeled PEDANTIC!?)

Re:3 ... 2 ... 1 ...

[dangitman]

I suppose that depends on which video cards [newegg.com] and SSDs [newegg.com] you use.

The problem with that SSD is that it is slow. One of the main benefits of SSDs is performance. If the SSD you buy doesn't give a marked increase in speed over a hard drive, then why pay 30x as much per GB for?

Re:3 ... 2 ... 1 ...

[Anonymous Coward]

The problem with SSDs is:
1. cheap
2. big
3. reliable
Choose two!
But even then, you can only be sure of number 3, after some years have passed. For obvious reasons of there not being any test data for years of use, until years of use have passed. ^^

ofc you mean cheap, small, reliable and only 2 of them can be obtained as in cheap, fast, quality

Re:3 ... 2 ... 1 ...

[Hurricane78]

Well, you can, but you would not define them as SSDs anymore, because they would not work at all, so they are not sold. Just go to search the trash cans of the SSD companies. ^^

Frankly, I don’t care about high performance or size. I still prefer big tower cases. (And yes, I can still fill them completely. ;)
What I care for is:
1. My data fits on there (everything above 1-2 TB).
2. It is 100% reliable. I don’t even touch any file system that isn’t a triple-mirrored ZFS with constant scrubbing, S.M.A.R.T. surveillance/testing and GIT-like “infinite undo” backup. I have lost half my data five years ago, and am still in the process of restoring it. Some of it will even never come back. Ever. Because it was so rare. (Vinyls that only 7 people on this planet ever had.) And: Yes, I already had a mirrored RAID and a reliable backup back then. But that won’t help you a bit with a controller that quietly writes data to the wrong sectors of your archive. Or degrading data, like with SSDs. So I won’t touch a SSD at all, until its degradation is like that of HDDs without all the trickery and magic that is required nowadays, to even get in the same league.
3. I can afford it. I don’t have much money, and making money equals lots of real pain and horrors for me at this time. So it’s either worth its price, or not. SSDs are not. Not even remotely. You pay 10 times more, but get 10 times less. No thanks.

Re:marketing dishonesty and abbreviations

[clone53421]

When you’re talking about storing (or transmitting) data bit-by-bit, it’s pretty common to see the rates being expressed in terms of bits. Terabits, gigabits per second, etc.

It’s slightly confusing at times, I’ll admit.

Another five-years-out technology

[alvinrod]

The technology sounds impressive, but then they just give it the kiss of death by announcing that it's five years away. Five years from now it will still be five years away, probably because while it's possible to do, no one has been able to do it in a cost-effective manner. Also if Intel can keep up with their current roadmap, they'll probably be using something close to a 10 nm process. I know that both Global Foundaries and TSMC are working on their 28 nm process (Although they are behind schedule.) so it's not inconceivable that the rest of the industry will already be at that point anyhow.

Re:Another five-years-out technology

[osgeek]

What's depressing is the way that the press and /. alike eat up stories like this.

Sure, writing this density of nanodots is an impressive feat. But as you point out, it could be completely nonviable for creating an actual consumer product.

Why can't the Slashdot's front page be the kind of place where bullshit is called on researchers putting out this kind of nonsense. These guys should be shamed into putting out factual press releases. Whoring it up to get coverage from the general media while seeking increased funding should be smited.

Here's a cool project: Go back into the Slashdot archives and pull out all the articles that made five-years-out predictions. Keep the oldest ones where the claim has not been fulfilled but the claimant is still in business/tenured sorted at the top of a slashbox that we can all jeer at.

Re:Another five-years-out technology

[canajin56]

I remember, like 15 years ago, a Byte Magazine special issue "The Future of Computing!" Said that 16 TB holographic CDs were only "5 years out" ;) We were also supposed to have algae-based RAM that was so large, and fast, plus non-volatile, so we wouldn't even need both an HD AND RAM, just one thing. Plus we'd also have quantum CPUs made out of doped diamond, that doesn't need a heatsink because diamond semiconductors don't get less efficient with heat, and can go a lot hotter without damage. AND HOVERCARS.

Jay Narayan works for NCSU, not UNC-CH

[Anonymous Coward]

Dr. Narayan teaches in the Department of Materials Science and Engineering at N.C. State, which *is* part of the UNC system of colleges and universities, but is a university in its own right, and is actually a fierce rival of UNC-CH. Here is his personal homepage at NCSU.

Re:Nanodots? ..b-bots?

[Anonymous Coward]

I read nanodots as nanobots...

Try reading a book once in a while, it helps improve your reading comprehension.

Dear Slashdot,


The jokes that go "I read something that was clearly not there so it doesn't say that but wouldn't it be funny/great/novel if it did, haha mod me up to +5 funny now kthx" got old a long time ago.

Single Defect Free Crystal

[kenp2002]

Ok my knee jerk Six-Sigma reflex has just kicked in. On the manfacturing of those defect-free crystals... and about the cost effect and scaling for "overtaking ... in 5 years..."

Ok, here is a tip:

Anytime a politician or scientist taks about 5,8, and 12 year targets there is a reason:

Two 4 year terms = 8 years; when the project falls out they can blame the canidate currently in office.

5 years = A single Term but just a touch beyond to provide an incentive for re-election because if you don't they might cancel the project

12 Year = Two terms for canidiate A and a term for his\her heir... "Don't let the evil Democrats\Republicans kill the project!"

Now last I checked more then a few grants come in at 3,5,8 and 12 year durations... I never hear things coming to fruition in 7 years, or 6 years, or 9 years, or 11 years, or 18 years, 6 months, and 3 days.

There is just something about 5, 8, and 12 they love. Which due to the frequency they cite those values implies there is some weird cosmic alignment which causes innovations to pop at those figures... or I smell 4/5 dentists approve BS.

Another one is the 20 years from now number. What is the maturity on that investment I made...

I would honestly have a lot more respect for senior scientists if they didn't spend every waking hour working on getting grant money leaving the actual work to low-paying interns and students then claiming the work as their own offering nothing more then a second hand "my team and I" comment...

No doubt..

[StoneOldman79]

I have no doubt that for The question will be if nanobot/SSD will beat hd's for large quantities of TB's per $ by 2015.
I'm afraid this will not be the case so for your 10TB+ video collection and big data centre stuff we will still be using hd's.

NANODOTS CAUSE IMPOTENCE

[Anonymous Coward]

Considering my investments in the current technology, let me be the first to start the inevitable FUD campaign. I'll even declare it upfront for everyone. NANODOTS PROVEN TO CAUSE IMPOTENCE. Discuss, share, let the fear spread. I'm fairly certain that they are also solely responsible for global warming. We had no warming until the DOTS!

and for the WoW crowd. STOP DOTS! STOP DOTS!

Re:NANODOTS CAUSE IMPOTENCE

[ProfMobius]

If so, I guess it will be labelled as such in California soon.

Access speed? Throughput?

[VincenzoRomano]

I wonder how do they think all those data can be made accessible with fast access speed and good throughput.
The article is failing to mention this slightly important topic. Also tapes can store a lot of data (well not really that lot) with ridiculous performances.

Does anyone "collect" porn anymore?

[sean.peters]

Given that there's an infinite supply of ever-changing pr0n on the internet available for free, I have to wonder why anyone would bother stashing it on a local disk. It's like saving bottles of air.

Re:Does anyone "collect" porn anymore?

[ProfMobius]

You don't save air in bottle ? When air will run out, I will be the one laughing !

Re:Does anyone "collect" porn anymore?

[Anonymous Coward]

Because it makes your penis bigger in more ways then one.

Re:Does anyone "collect" porn anymore?

[An ominous Cow art]

I save time in a bottle - you never want to run out of time.

Then there's the "Friedman unit"

[sean.peters]

... aka six months. For those just tuning in at home, this unit of time was popularized by one Thomas Friedman, a columnist who was thoroughly mocked on the Internets for his unfortunate habit of claiming that we needed another six months in Iraq to know if it was a success, and then when the six months had gone by, proclaim that another six months was needed. Over and over and over again - to the point where various critics would make a note on their calendars, and then after six months ask "well, it's been another Friedman unit... can we go home now?" Would have been funny, except that it was sad.

TFS is wrong; it's North Carolina State University

[osvenskan]

TFS gets it wrong; this research was done at North Carolina State University (NCSU), not University of North Carolina (UNC).

Here's the NSCU press release:

http://news.ncsu.edu/releases/degraffnarayan09/

UNC? The man is at NCSU.

[Anonymous Coward]

The Dr. in question is a joint professor at UNC (The cuddly little school where all the journalism majors come from) and at NCSU (The campus covered in bricks that churns out engineers, and where the good doctor happens to be chair of microelectronics). Which school do you think the research was done at?

Lame Research?

[GameGod0]

It may be peaking soon though. 6nm is getting close to physical maximums for most techniques due to the casimir effect.

Not quite sure what the Casimir Effect has to do with magnetic dots, but I should mention that 6 nm is below the Superparamagnetic limit (which is typically tens of nanometers). That means you're magnetic nanodot probably isn't magnetic.

... Which brings me to my second point: This article says nothing about what this researcher actually did. It sounds like he just fabricated an array of nanodots, which is nothing particularly groundbreaking.

Does anyone have a link to the original abstract for the conference presentation? The dots must have been multilayer "stacks", otherwise there's a good chance they won't be ferromagnetic (there's a "superparamagnetic limit" that stops ferromagnetic particles from being ferromagnetic when they get around this size.)

Lastly, the article says they'll look at housing and using "laser technology" to read back from these nanodots. They mention that as a sidenote, but it's really the most important problem if you want to make something useful. The problem with most nanomagnetic memory techniques is that reading/writing is either impractical or not yet possible.

Re:Lame Research?

[HotNeedleOfInquiry]

... Which brings me to my second point: This article says nothing about what this researcher actually did. It sounds like he just fabricated an array of nanodots, which is nothing particularly groundbreaking.

Exactly. And I had to wade through about 10 pages of silly off-topic comments to find confirmation of that. Thank you Game, thank you lame moderators.

Re:Lame Research?

[rrohbeck]

The Casimir effect has to be considered for the flying height budget in hard drives, FWIW. Magnetic dots are all the rage for hard drives too ("patterened media.")
But for solid state? WTF? And how is laser light supposed to resolve 10 nm dots?

Re:Lame Research?

[Idiomatick]

Yeah, I wasn't sure what specific technique he was using but since he will be running into fundamental physical limits either way it still stands.

It does sound like lab only uselessness that can't be used for storing data. Not just impractical, but doesn't work as data storage (readable/writeable with some degree of accuracy).

Re:Lame Research?

[Anonymous Coward]

Superparamagnetic structures are still magnetic and, depending on the material, might be ferromagnetic. If we define ferromagnetism as locally parallel aligned atomic magnetic moments, that is.
The atomic moments are still aligned in the superparamagnetic , it is just that the direction of the total magnetic moment (sum of all atomic moments) becomes unstable. Basically, below a certain size a dot will switch back and forth between semi-stable states.
This is obviously unusable for information storage, but you can still do some funny physics with it, if you can control the properties of the semi-stable states (which, incidentially, we can).

IAAPINM (physicist in nanomagnetism)

(No idea why GP was referring to the Casimir effect, either...)

Re:Nanodots? ..b-bots?

[Anonymous Coward]

I read Tb as TB. because 4Tb sounds more than 512MB. /me rages hard against this intentional misrepresentation and concludes with apathy

It's 500GB

[xaxa]

4 Terabits = 512 Gigabytes

Except it doesn't.
4 Tb = 4 000 000 000 000 / 8 B = 500 000 000 000 B = 500 GB ~= 466 GiB

Did they mean 4 Tib?
4 398 046 511 104 / 8 B = 549 755 813 888 B = 512 GiB ~= 550 GB.

According to the scientist, it's the former:

"at 10nm per bit, 1cm square stores one terabit."

That would be (1cm / 10nm)^2 b = (1e-2 / 1e-8)^2 b = 1e12 b = 1 Tb.

Think different

[Anonymous Coward]

What if that kind of thinking was pervasive in all branches of technology?

Why make crops yield better, as we can simply use more land? Why advance medicine as we can just kill more patients?

Point is, you don't always have to use more resources to achieve your goal. The alternative is to use the resources you have more efficiently.

Sure, RAM is cheap. But without that mindset we would have computers that are truly thousands of times faster than what we have today.

Maybe you are satisfied with mediocrity. I'm not.

Re:Think different

[Jbcarpen]

There is a limit to Moore's law, there has to be. and when we reach that limit we'll see focus be put back on optimization rather than ever faster hardware. Yes, it'd be much better to shift focus sooner, but I doubt we'll see it happen.

Re:Think different

[Michael Kristopeit]

There is a limit to Moore's law, there has to be.

you don't sound like you're convinced...

NCSU not UNC

[Anonymous Coward]

Narayan is located at North Carolina State University, not the University of North Carolina! Big difference if you went to either school ;)

Why can't we have more? Conversation below

[Merc248]

Scientist: "Dots! More dots!" ...

Scientist: "Okay, stop dots."

I agree...

[crhylove]

With your complaints. So let's start a list of UNbloated software:

I'll start:
MicroXP.

Why software bloats

[aminorex]

Bad software managers are rewarded for producing a lot of software. The more software, the more reward. As a result, you get increasingly useless or downright harmful crap rammed down your throat whenever you buy a commercial software product or a piece of hardware with bundled software. The latter is the worst, because in the case of commercial software there is at least a reality check which comes from the need to prevent the product from becoming so odious that no one will buy it.

You have the wrong university...

[Anonymous Coward]

Wrong University - North Carolina State University (NCSU) vs. University of North Carolina (UNC).

Delicious

[StikyPad]

If nanodots are anything like Dippin' Dots, they sound delicious.

Re:TFS is wrong; it's North Carolina State Univers

[John Hasler]

The abstract is not very promising.

anonymous coward

[Anonymous Coward]

The article states that 4 TB is 512 GB. As far as I know, 1 TB is 1024 GB, making the chip 4096 GB. Is this incorrect?

Re:anonymous coward

[lyml]

It's terabits, one byte is 8 bits thus the confusion.

It's NCSU....

[nicks,nicks,nicks!]

It's North Carolina State University, not University of North Carolina.