Titan, World's Fastest Supercomputer (20 Petaflops), page 1

Pages: << 1 2 >>
ATS Members have flagged this thread 7 times

Topic started on 1-11-2012 @ 11:23 PM by Raelsatu

Yet another supercomputer. Before 2020 multiple countries have stated exascale computing.

DoE Activates Titan

When introducing Titan to the world, the U.S. Department of Energy's (DOE) Oak Ridge National Laboratory said it is “the world’s most powerful supercomputer for open science with a theoretical peak performance exceeding 20 petaflops (quadrillion calculations per second).” That almost unimaginable computational capability is like if each of the world’s 7 billion people were to solve 3 million math problems per second. To better help you wrap your mind around just how fast Titan is, “It would take 60,000 years for 1,000 people working at a rate of one calculation per second to complete the number of calculations that Titan can process in a single second,” explained National Geographic.

“The Cray XK7 system contains 18,688 nodes, with each holding a 16-core AMD Opteron 6274 processor and an NVIDIA Tesla K20 graphics processing unit (GPU) accelerator. Titan also has more than 700 terabytes of memory.”

Titan’s supercomputer predecessor will be an exascale

The next ORNL supercomputer is to be an exascale, meaning it will be a thousand times more powerful than Titan and capable of doing one quintillion calculations per second. Exascale power is supposed to “provide enough power to simulate every single atom in a whole living cell,” Smith explained.

edit on 1-11-2012 by Raelsatu because: (no reason given)

reply posted on 1-11-2012 @ 11:59 PM by YoureAllNuts

but does it play minesweeper?

reply posted on 2-11-2012 @ 12:14 AM by OneLove20

What is it going to be used for?

reply posted on 2-11-2012 @ 12:20 AM by minkmouse

GIGO at hyper speed. It's truly awesome how far we've come with computational power but can we keep up, I mean you could give a Chimpanzee an etch a sketch but can he draw a banana? I fear the main use we'll find for this may not take us places we want to go. Damn that's fast!

edit on 2-11-2012 by minkmouse because: (no reason given)

reply posted on 2-11-2012 @ 12:22 AM by ChaoticOrder

reply to post by Raelsatu

The next ORNL supercomputer is to be an exascale, meaning it will be a thousand times more powerful than Titan and capable of doing one quintillion calculations per second. Exascale power is supposed to “provide enough power to simulate every single atom in a whole living cell,” Smith explained.

lol so a computer 1000x more powerful than Titan would still only be able to simulate a single cell with real world atomic physics. Imagine trying to simulate our solar system or the observable Universe.

reply posted on 2-11-2012 @ 12:26 AM by MastaShake

Originally posted by ChaoticOrder
reply to post by Raelsatu

The next ORNL supercomputer is to be an exascale, meaning it will be a thousand times more powerful than Titan and capable of doing one quintillion calculations per second. Exascale power is supposed to “provide enough power to simulate every single atom in a whole living cell,” Smith explained.

lol so a computer 1000x more powerful than Titan would still only be able to simulate a single cell with real world atomic physics. Imagine trying to simulate our solar system or the observable Universe.

with how fast everything is advancing i wouldnt be surprised if we were able to do that in the next 25-50 years

reply posted on 2-11-2012 @ 12:29 AM by minkmouse

reply to post by ChaoticOrder

Just give it a few years, maybe say 2045? We've come from univac to here in sixty years and change and it was just yesterday we were all tipsy with "Pac Man"

reply posted on 2-11-2012 @ 12:30 AM by ChaoticOrder

reply to post by MastaShake

We will have a chance of being able to do it in the next 25 years if we develop proper quantum computers. Otherwise we are a lot further off than 25 years.

edit on 2/11/2012 by ChaoticOrder because: (no reason given)

reply posted on 2-11-2012 @ 12:37 AM by SaturnFX

Human brain estimates - 100 petaflops

Now, if we consider how much redundancy, ineffective wiring, etc the human brain has, I would venture to say this machine already equals of not surpasses us (if it had the proper programming installed and a good medium).

Of course shrinking this down from room sized to grapefruit or smaller would be helpful to get it into a mobile unit.

Time to start considering Asimov's 3 laws of robotics

1.A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2.A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
3.A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

I think I would add a few more things in there..such as, if a human tells you to go break into that bank for them..nix that command.

Actually, I am more interested in the time when we can take this processor, adapt it to interface with our brain, and have mega processing power at will for our own selves.

edit on 2-11-2012 by SaturnFX because: added stuffs

reply posted on 2-11-2012 @ 12:53 AM by SaturnFX

Originally posted by ChaoticOrder
reply to post by MastaShake

We will have a chance of being able to do it in the next 25 years if we develop proper quantum computers. Otherwise we are a lot further off than 25 years.
edit on 2/11/2012 by ChaoticOrder because: (no reason given)

I think your mistaken. in 25 years, the computer industry can grow in massive leaps.
Consider 25 years ago. 1987. the best computer on the market was the IBM PS2 8580 that ran upwards of 10 thousand dollars or more
and it had a 20 MHz processor 386. aka, it ran dos and some general crap stuff.

25 years from now, our modern warfare or skyrim will seem like pong...its almost unimaginable without looking at star trek's holodeck for reference of potential..and hell, that might even seem quaint. There are many paths towards massive acceleration of computational powers. Quantum is one aspect already being employed.

in 25 years, I expect claytronics (programmable matter). to be the big thing

reply posted on 2-11-2012 @ 01:05 AM by liejunkie01

It still amazes me that some people do not think that tptb can monitor every thing you say on a phone or type on a keyboard.

They use that computing power for something folks and guess what?

This is not the only super computer out there.

Why would the department of energy want or need that kind of computing power.

And think about the computing power when it is linked to other super computers.

Simply mind boggling.

reply posted on 2-11-2012 @ 01:18 AM by ChaoticOrder

reply to post by SaturnFX

I know exactly what you are saying, but I don't think you quite comprehend the difference between a single cell and an entire solar system, let alone the entire observable Universe. If we can simulate just this solar system within 50 years I will be extremely impressed.

Think about how huge these super-computers must be to simulate just a single cell. We are using these massive machines which take up large rooms just to simulate something which can't even be seen with the naked eye.

In fact they haven't even built a computer powerful enough to simulate a cell yet. Titan is the worlds fastest but is still a thousand times slower than it needs to be. It seems to me that if we wanted to simulate this entire Solar System we would most likely need a computer larger than the solar system.

Remember that each and every atom in the solar system represents a piece of information which needs to be simulated. With classical computers we will not simulate the entire solar system any time soon. Unless of course we develop powerful quantum computers.

reply posted on 2-11-2012 @ 03:12 AM by C0bzz

177 trillion transistors is what got me.

Also a better article is found here:

www.anandtech.com...

reply posted on 2-11-2012 @ 03:23 AM by Raelsatu

Originally posted by ChaoticOrder

lol so a computer 1000x more powerful than Titan would still only be able to simulate a single cell with real world atomic physics. Imagine trying to simulate our solar system or the observable Universe.

Yes, well there are a few factors that come into play here. It may take an exaflop to simulate a single cell; it also depends on how accurate you want the simulation to be. If they plan on making a simulation that mimics our universe & its functions down to the sub-atomic scale, quantum forces, laws of physics, etc -- well then yes the computational power required is enormous.

In India, ISRO and Indian Institute of Science have stated that they have planned to make a 132.8 exaflop supercomputer by 2017

They're claiming a 132 exaflop system within 5 years, which seems incredibly soon, yet impressive if they can pull it off. 2030 is the estimated time by which we'll have zettaflop supercomputers

So we're talking about classical computing. But what about quantum? There's a new Princeton article out showing research done that lays out a path for quantum computers with millions of q-bits. Which is unprecedented.... until now scientist have only been working with q-bits in the 10s or 100s. The implications here appear to be huge.

reply posted on 2-11-2012 @ 07:46 AM by lasertaglover

Skynet!

Where is John Connor at?

reply posted on 2-11-2012 @ 08:08 AM by Cauliflower

This looks like some kind of disclosure rather than an RFC doesn't look like complete disinfo at least in theme.
Maybe they are working with AMD because of an optical Scyld Beowulf 64-Bit interface?

I only scanned the earlier replies briefly, there were 12 Titan children of Ouranos and Gaea, some took up arms against Cronus yada yada.

reply posted on 2-11-2012 @ 08:26 AM by moniesisfun

reply to post by ChaoticOrder

No, I don't think you grasp what Sfx is saying at all. Sorry.

Organic and quantum computing is a whole new ballpark. Things that are processed incredibly inefficiently through silicon systems the size of a stadium today, might take one cycle of one chip within a decade or so. That's only the stuff we can currently imagine.

reply posted on 2-11-2012 @ 12:04 PM by ChaoticOrder

reply to post by moniesisfun

Organic and quantum computing is a whole new ballpark.

Obviously you didn't even read the part where I mentioned quantum computers (although it's still not very clear whether practical quantum computers are even possible in the way we imagine). I'm saying classic computers will never do it.

reply posted on 2-11-2012 @ 04:05 PM by moniesisfun

reply to post by ChaoticOrder

Apologies.

Rough week.

reply posted on 4-11-2012 @ 10:58 AM by PlanetaryDuality

reply to post by OneLove20

“the world’s most powerful supercomputer for open science"

Line two.

Pages: << 1 2 >> ^^TOP^^