Computer Capable Of Calculating 640TBs Of Data In One Billionth Of A Second (available to public by

originally posted by: DigitalJedi805

In the early 90s, nobody ever thought anyone would fill up a hundred Megabytes. It's not likely that anyone will ever fill that much data with books or notes; but as multimedia grows, so do the space requirements. The company I work for ( contrary to my advisement ) still encodes video at 640x480, and five hours will generate a gigabyte of data. If we were to double that ( which we very well should, as well as moving to widescreen... ), math would dictate that we would end up with two gigs of data for the same information. Considering that resolutions are growing Far beyond what I'm talking about here, when you start dealing with things like... For instance... The 24 hours of Le Mans; if it were to be recorded at 20000 frames at 2560 resolution ( like it should damnit! ), I'd bet you're looking at terrabytes of data. That's just one example.

I work for a marketing / IT company and as part of that, we have a video studio for shooting commercials and things of that sort. we also have a 3d rendering farm for animation.

At the core of all of that is a 12 terabyte file server.. for one project that has 5,460 frames ( 3m, 45s video ) .. each frame was about 3.5mb in size at 1920x1080 resolution .. the frames folder is just shy of 20 gigs for 3 1/2 minutes of animation.

a reply to: miniatus

12 TB doesn't seem like that much, considering I have a 1 TB external HD for all my media that's the size of a deck of cards.

I would imagine doing the kind of work you do, you'd have 200 TB+ to work with?

a reply to: Grimpachi

I don't know why more people are not ex-static about this this is revolutionary technology all of are current technology for the general public right now is now a fossil in comparison. That's how big this is I mean come on were talking Peta bytes of data in nano seconds the photon technology in this thing alone is amazing.

It will take something like this to really take computer technology to the next level. One could argue that a person would never need this much power, but it must be realized that more and more complicated programming will be performed, and we will have software in the future that can do crazy things, and will put that extra computing power to good use. What I want to know is what this means for the future of gaming? Haha. Honestly though, there will always be drawbacks. Even now we are complaining about government oversight, and they will have this technology as well. We think the mass surveillance is bad now? With the kind of power we are talking about with a network built around devices like this, I shudder to think of the possibilities for clandestine activity by the government against its own citizens. The ability to create programs to process the data will improve as well, and who knows what this could mean?

That's good, when we are ready for warp drive engines we will need that kind of computing power on the ship.

a reply to: Grimpachi
Very cool, but what there doing has been a project for a long time its not necessarily new more like coming to a market next to you. Basically there taking the fastest computing thing known to man and thinking on how it operates, then reworking and copying that into a silicon synthetic materials ie computer. The fastest computing thing being the human brain, computer may be able to function at one or a few things very fast but if they try to operate at an mass amount of operations at the same time they break down, its why your computer slows down when its inflicted with viruses or even your running more programs on it then it can handle, the human brain layout being more malleable for such thing it can operate different things all at the same time due to its structure and how its set up.

Its pretty cool I suppose, and who know one day we may have computers and AI become sentient, hey if it can store the entire life a person on a phone with this tech like they said, it already has the some of the things required. Though there is that whole thing of making humans obsolete thing, could be that like in the anime Ghost in the Machine future flesh and blood humans are mostly living in outdated slum cities, and anybody who is anybody is mostly or all synthetic in body and with a whole half biological half synthetic cyberbrain. Oh ya the future is looking more interesting, though it can also be looked at as a form of hell, especially for those who are part of the outdated past. Two edged sword indeed as this becomes more and more a reality due to technology. Though so far it will just mean faster and smaller computers and gadgets.

It'll only be popular if people can put selfies on Facebook and find amusing photos of cats with it.

Think Star Trek tri-corders. Someone else mentioned teleportation. DNA sequencing, a myriad of new possibilities. We want to explore space at near or light speed? Going to have to be able to crunch a lot of data to ensure you arrive at your destination safely. Global weather patterns / changes, earthquake predictions. You name it. Suddenly, those things aren't so far fetched.

It does seem pretty surreal to be sure, but it has to start somewhere doesn't it?

originally posted by: JJRichey
a reply to: Grimpachi

*Disclaimer* I make no claims to be an expert in this field, but as someone who took a few semesters of computer science, this is my viewpoint on this. I could be totally wrong in my understanding, wouldn't be the first time!

Thats the beauty of this technology. It uses such little power that your liquid cooling setup wouldn't be needed. After all, the majority of your computer systems heat is generated from all those electrical components heating up under stress. Using photonics should generate next to zero heat, as opposed to using copper traces and electricity like is used now. This isn't talking about just a new computer technology, its talking about completely changing the way computers work.

I wonder if they can borrow "duotronic" from the Roddenberry estate.

en.memory-alpha.org...

But seriously it really does appear to be the biggest change in fundamental hardware representation since the 60's and the integrated circuit.

This technology is marketing nonsense.

Memristors are certainly in the near future, and they will certainly boost computing performance by an order of magnitude. But the CPU's HP are talking about are nonsensical. The figures they have produces are based of the theoretical throughput of the memory. However...

By their own admission, the CPU will be comprised of many, many different core specialising in different functions. The problem is, is that there are no real world applications that require many types of different processing. If you need true serial processing (calc.exe for example) you'll be limited by the fact that a single 'core' is terribly slow. Certain high end applications (particularly those that deal with 3D modelling) will only access a small percentage of cores too. Which kinda defeats this type of computing - you pay for this much capability, but you'll only use 1% of it!

Having said that, the power reductions will be attractive to many data centres/render farms in the near term.

a reply to: MarsIsRed

Did you watch the video? The reason I ask is because you are claiming there are no real world applications but he listed some real world applications. Try to remember this tech will be released to businesses many years before it hits the general public just like almost all things such as this. In that time I am sure there will be plenty of real world applications for it because programers will no longer be constrained to the current mainframes.

The limitations to this seem to be just short of ones imagination and programing ability.

a reply to: Grimpachi

Yes, I watched the video!

What HP are suggesting is moving away from ARM vs Intel vs GPU. and introducing more types of processor. This ONLY works if you have a great many jobs concurrently. Without the concurrency you have limited computational ability.


I'm not saying this can' work, it's just implausible.

a reply to: Grimpachi

what is this?

A Magnetic Spin switch?

Optical logic switches?

Quantum dot switches?

All of which are millions(potentially billions) of times faster than normal CMOS/silicon switches we have now.

a reply to: MarsIsRed

Doesn't that make this great for servers compared to what we have now?

a reply to: Grimpachi

Can they make this in mass volumes at dirt cheap prices?

That is the question.

a reply to: John_Rodger_Cornman

If it is like all other hardware it will probably be outrageously expensive at first but within a few years it will be affordable a few years later it will be cheap and integrated into everything.

Then again I could be wrong because they are cutting down on components and the introduction price to the general public could be moderate. I just can't say I can only guess. I do know that the more the private sector adopts the tech and the need arises for more manufacturing plants the price should drop significantly with each as they refine their manufacturing process.

I hope they keep their 2020 timeline for average joe release. Depending on the price I may actually wait in line for these. Otherwise, I will use the best of the older tech for a few years waiting for this to become affordable.

a reply to: Grimpachi

ok

This will be used heavily by google,DoD,physics simulations,protein folding, and colleges.

originally posted by: surfinguru
Think Star Trek tri-corders. Someone else mentioned teleportation. DNA sequencing, a myriad of new possibilities. We want to explore space at near or light speed? Going to have to be able to crunch a lot of data to ensure you arrive at your destination safely. Global weather patterns / changes, earthquake predictions. You name it. Suddenly, those things aren't so far fetched.

It does seem pretty surreal to be sure, but it has to start somewhere doesn't it?

All true...
And as i said in my prev post:
The list for potential uses
will be almost endless...

All this incredibly smart science and technology talk and Meg Whitman had to slaughter her credibility by using the term drownding @28:39. Get off the stage lady, you have no idea what is going on.

It is pretty amazing that the term petabyte will be very common in the near future. memory in the terabyte level? that is going to be incredible.

Electrons compute, photons communicate, and ions store. I was so excited to hear about graphene and the potential for fast data transfer, but this seems to blow it out of the water. Graphene for batteries, and this technology for computing, the power seems limitless.

An open source cloud environment? sounds fancy but also dangerous. He was talking about making airline data anonymous over the cloud, but leave it to a hacker and this data can be revealed. Cyber criminals are going to hit an all time high soon.

But with less energy and insane speeds, the networks of tomorrow are going to be cheap and faster than we will ever need.

the future is scary and enticing, i cannot wait to see what happens next. time to invest some stock into HP and hope for the best.


a reply to: Grimpachi

a reply to: Aliensun

TB (Terra Byte) is a measure of storage, not power. Carrying all of your data in your pocket would be possible. Storing not a few, but many movies on your phone would be cake, and you'd still have room for more data on top of that. When storage is no longer a problem, many new types of data will be stored. Forget watching low-res videos when you have the storage space to keep hi-def videos.

On top of the mega-storage capabilities, they claim to have a processing architecture that allows the data to be processed extremely quickly with a fraction of the power requirements. That's interesting.

The big problem with devices of this capacity is transferring the data to and from them. We're still limited to transfer rates of our existing infrastructure. Maybe this increase in processing power, we can come up with some better models to allow faster transfer of this new data capacity.