![Memory [Forever]: Bits Never Die](http://cache-06.gawkerassets.com/assets/images/4/2010/03/126x100_memory-forever-1.jpg "Memory [Forever]: Bits Never Die")
Enter your username and password.
" 1.74 quadrillion floating points operation" <----- My response, "Holy shit!"
Following Moore's law, can you imagine in 50 years when personal computers will have that computing power. At such computing power will all PC's be identical having capped the markets need for computing? What would an individual need with such computing power? #jaguar Reply
Following Moore's law, can you imagine in 50 years when personal computers will have that computing power. At such computing power will all PC's be identical having capped the markets need for computing? What would an individual need with such computing power? #jaguar Reply
@Michai: until we have robots that can make me a sandwich, there will always be a need for more computing power. #jaguar
Reply
@Michai: I always find this helpful to remember regarding Moore's, or any exponential growth situation really:
[en.wikipedia.org]
On 13 April 2005, Gordon Moore stated in an interview that the law cannot be sustained indefinitely: "It can't continue forever. The nature of exponentials is that you push them out and eventually disaster happens." #jaguar Reply
[en.wikipedia.org]
On 13 April 2005, Gordon Moore stated in an interview that the law cannot be sustained indefinitely: "It can't continue forever. The nature of exponentials is that you push them out and eventually disaster happens." #jaguar Reply
@NorwoodIsMyHero: "...eventually disaster happens."... where a new exponential curve begins. #jaguar
Reply
@Michai: Then in 52 years this will be the power of a celeron and for some reason even though this much power was no where near available 4 years before you will feel the need to the fastest newest tech to check your email on. #jaguar
Reply
@Michai: Haha, so true, I find it hard to imagine what disaster would befall us if transistors kept getting exponentially smaller.
He did also point out that their are ultimate limits at the atomic level to how small things can get. #jaguar Reply
He did also point out that their are ultimate limits at the atomic level to how small things can get. #jaguar Reply
@Michai: I'd argue closer to 25 years rather than 50.
According to this article ([www.macobserver.com]) the fastest Mac Pro in 2008 was 1/12,142 as fast as RoadRunner.
Given that ratio, it will only take 14 doublings (which is 16,384 times more) to reach this speed for a consumer computer.
Moore's law says we should expect a doubling of transistors on a chip every two years which comes to only 28 years.
Ray Kurzweil argues (convincingly) that the doubles are occurring faster than every two years and in fact are themselves speeding up. So 28 years is probably a very conservative estimate.
But will it run Crysis? #jaguar Reply
According to this article ([www.macobserver.com]) the fastest Mac Pro in 2008 was 1/12,142 as fast as RoadRunner.
Given that ratio, it will only take 14 doublings (which is 16,384 times more) to reach this speed for a consumer computer.
Moore's law says we should expect a doubling of transistors on a chip every two years which comes to only 28 years.
Ray Kurzweil argues (convincingly) that the doubles are occurring faster than every two years and in fact are themselves speeding up. So 28 years is probably a very conservative estimate.
But will it run Crysis? #jaguar Reply
@Michai: Since we've made similar statements for decades now while Moore's Law has been working and had all those needs filled and more, I'm believing that you'll need this power for your latest 3 dimensional Hollo-Deck display in the game room of your house by then where a MMG finally exceeds 100 million simultaneous players and even the cockroaches walking under your feet are being modeled and you have infinite zoom to be able to check molecules in objects. Bottom line, whatever computing power is created will cause applications to be written that will enable it's use, and more... #jaguar
Reply
@kosai: Thanks for the vid, very cool. I'd like to add that it does not account for the intelligence of software aiding in the conductibility of trillions of computers. #jaguar
Reply
@NorwoodIsMyHero: Here is a cool fact that everyone may find interesting. All of todays digital information, if translated into a physical weight, would weight about the same as large grain of salt*. Given that knowledge a atomic matrix the size of a 6 sided die has the potential to hold more than enough information for generations to come.
*(I read this story in either one of my Scientific American or Discover magazines, I spent a few minutes looking for the article but could find which month or magazine it was. Props to the person who locates the original article for me. It was within the last 9 months.) #jaguar Reply
*(I read this story in either one of my Scientific American or Discover magazines, I spent a few minutes looking for the article but could find which month or magazine it was. Props to the person who locates the original article for me. It was within the last 9 months.) #jaguar Reply
@SkipErnst: Yes true, but Kurzweil doesn't account for - like anything the closer you get to reaching the exponential's vertical the longer it takes. Also a change in transistor tech from silicon to 'whatever' will cause an initial slowing of the curve. #jaguar
Reply
@brianmi40: Your probably right. Can you imagine programming for 240bit software platforms!? #jaguar
Reply
@Michai: I seem to recall reading somwehere that every year (or for the past few years, or in one particular year, or something equally similar) we created as much digital information as already existed. Yeah, my memory is extremely fuzzy on this "fact".
But if it's even close we're gonna need more than one of those 6-sided dies to store all the info generated in the generations to come. #jaguar Reply
But if it's even close we're gonna need more than one of those 6-sided dies to store all the info generated in the generations to come. #jaguar Reply
@Michai: When I can render Crysis in real time, inside of a virtual machine, while editing pictures in PhotoShop, and watching 7 YouTube videos in HD at the same time, we can stop making faster computers. #jaguar
Reply
@SkipErnst: The problem is that is that silicon semiconductor fabrication is going to run into some very real physical limits in the next 10-15 years. Somewhere in the single digit nanometer levels, quantum effects will increase the noise level to the point that you cannot reliably function. Flash memory is expected to hit a limit much sooner. At that point, you can either start geometrically expanding vertically, or come up with some radically new tech path. #jaguar
Reply