dyask
|
|
July 07, 2014, 10:24:15 PM Last edit: July 12, 2014, 10:48:03 PM by dyask |
|
I don't think we will see 14 nm asics very soon. Becaus rhe smallest process node currently is 20 nm. Only Intel has smaller ones at the moment.
Intel does produce 14 nm chips, but I haven't heard of 14 nm ASICs yet. There are 14 nm FPGAs, http://www.altera.com/devices/fpga/stratix-fpgas/stratix10/stx10-index.jspThat still doesn't change my view. It won't be some guys taking about problems in a bar over nachos that end up delivering 14 nm ASICs. Also when we do see them, they won't be the huge jump in performance that might be expected.
|
|
|
|
|
|
|
|
"This isn't the kind of software where we can leave so many unresolved bugs that we need a tracker for them." -- Satoshi
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
|
|
|
|
dyask
|
|
July 12, 2014, 10:49:03 PM |
|
It isn't surprising that this is a scam.
|
|
|
|
ranochigo
Legendary
Offline
Activity: 2954
Merit: 4166
|
|
July 13, 2014, 01:28:42 AM |
|
Impossible, the smallest nanometer is 20nm, there wouldn't be such a decrease so soon. I would be interested to see his actual working product if he claim its 14nm. I don't think we can hit 14nm within a year
|
. .HUGE. | | | | | | █▀▀▀▀ █ █ █ █ █ █ █ █ █ █ █ █▄▄▄▄ | ▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀ . CASINO & SPORTSBOOK ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ | ▀▀▀▀█ █ █ █ █ █ █ █ █ █ █ █ ▄▄▄▄█ | | |
|
|
|
dyask
|
|
July 13, 2014, 04:17:18 AM |
|
Impossible, the smallest nanometer is 20nm, there wouldn't be such a decrease so soon. I would be interested to see his actual working product if he claim its 14nm. I don't think we can hit 14nm within a year
Actually the first 14nm fpga chips started showing up summer of 2013. Well better than in power consumption and the amount of gates available, they aren't vastly better though. http://www.fpgacentral.com/news/2013/altera-14nm-stratix-and-20nm-arria-fpgaHowever, I haven't heard of any 14nm ASICs yet.
|
|
|
|
ALToids
|
|
July 13, 2014, 06:23:16 AM |
|
Impossible, the smallest nanometer is 20nm, there wouldn't be such a decrease so soon. I would be interested to see his actual working product if he claim its 14nm. I don't think we can hit 14nm within a year
Actually the first 14nm fpga chips started showing up summer of 2013. Well better than in power consumption and the amount of gates available, they aren't vastly better though. http://www.fpgacentral.com/news/2013/altera-14nm-stratix-and-20nm-arria-fpgaHowever, I haven't heard of any 14nm ASICs yet. There's no reason to jump the gun to a more difficult and costly architecture when the "older" stuff still hasn't been optimized. Some of the 20/28nm stuff is not as efficient as it could/should be. Now if 14nm was cheap that would be another story...
|
|
|
|
dyask
|
|
July 13, 2014, 06:34:25 AM |
|
Impossible, the smallest nanometer is 20nm, there wouldn't be such a decrease so soon. I would be interested to see his actual working product if he claim its 14nm. I don't think we can hit 14nm within a year
Actually the first 14nm fpga chips started showing up summer of 2013. Well better than in power consumption and the amount of gates available, they aren't vastly better though. http://www.fpgacentral.com/news/2013/altera-14nm-stratix-and-20nm-arria-fpgaHowever, I haven't heard of any 14nm ASICs yet. There's no reason to jump the gun to a more difficult and costly architecture when the "older" stuff still hasn't been optimized. Some of the 20/28nm stuff is not as efficient as it could/should be. Now if 14nm was cheap that would be another story... 14nm isn't cheap and it requires a lot of redesign to get to the 2x performance level. You can bet that random guys jawing in a bar are able to do that level of design.
|
|
|
|
R4v37
|
|
July 13, 2014, 07:18:25 AM |
|
new generation = higher hash rate = higher price = Wont ROI = Diff. goin crazy
am i supposed to say this 14nm is impossible for now? i mean, it wont sell like hotcake
|
|
|
|
Zawamiya
|
|
July 13, 2014, 07:21:15 AM |
|
14nm is coming to intel, it might take awhile before 14nm used for ASIC.
|
|
|
|
dyask
|
|
July 13, 2014, 08:55:20 AM |
|
14nm is coming to intel, it might take awhile before 14nm used for ASIC.
Actually it is Intel currently producing 14nm chips. However, you are correct it may be some time before we see 14nm ASICs.
|
|
|
|
|
Domino
|
|
July 13, 2014, 09:05:21 AM |
|
Well either the power becomes a major factor or if the growth is slow then the 2016 halving will stop it.
Yup, once the halving happens it will be a whole new game. Do we know when exactly when halving will occur at this rate? The next halving will occur at the 420,000 th block, which will be found at around Aug 2016 according to http://bitcoinclock.com/.
|
|
|
|
Nagle
Legendary
Offline
Activity: 1204
Merit: 1000
|
|
July 15, 2014, 06:09:11 AM |
|
As for the original post, scam.
It's one of those "flexible funding" Indiegogo projects, which means "even if we don't get funded, we keep your money".
|
|
|
|
Bitsaurus
|
|
July 15, 2014, 09:07:41 AM |
|
Well either the power becomes a major factor or if the growth is slow then the 2016 halving will stop it.
Yup, once the halving happens it will be a whole new game. Do we know when exactly when halving will occur at this rate? The next halving will occur at the 420,000 th block, which will be found at around Aug 2016 according to http://bitcoinclock.com/. It will stay at August if there's no growth in the network. If there's a decrease in the difficulty it will get pushed back, if increase it moves up. It will probably be sometime between May and June 2016.
|
|
|
|
ThomasCrowne
Full Member
Offline
Activity: 210
Merit: 100
★☆★ 777Coin - The Exciting Bitco
|
|
July 16, 2014, 01:51:53 AM |
|
Moore's law is approaching it's end. We won't keep seeing leaps in transistor reduction. Switching to graphene or some other exotic manufacturing might be the future.
But for what seems an eternity in the Bitcoin world, the next 1-2 years will still see more efficient miners released.
Agreed. Seriously doubt asic miniaturization will slow down within the next 18-24 months, then after that I suspect we will see a slew of optimization techniques that carry this trend on for another 6-12 months easy. Of course, all of this is largely contingent upon the bitcoin price continuing it's upward trend.
|
|
|
|
Blockhunter
|
|
July 18, 2014, 07:51:32 AM |
|
As for the original post, scam.
It's one of those "flexible funding" Indiegogo projects, which means "even if we don't get funded, we keep your money".
bump!
|
|
|
|
DrG
Legendary
Offline
Activity: 2086
Merit: 1035
|
|
July 18, 2014, 09:03:06 AM |
|
Moore's law is approaching it's end. We won't keep seeing leaps in transistor reduction. Switching to graphene or some other exotic manufacturing might be the future.
But for what seems an eternity in the Bitcoin world, the next 1-2 years will still see more efficient miners released.
Agreed. Seriously doubt asic miniaturization will slow down within the next 18-24 months, then after that I suspect we will see a slew of optimization techniques that carry this trend on for another 6-12 months easy. Of course, all of this is largely contingent upon the bitcoin price continuing it's upward trend. I believe ALToids was saying the opposite of what you're saying. CPUs have pretty much reached a wall. ASICs for Bitcoin still have an efficient jump to 20nm, but 14nm will be cost prohibitive because all the foundries will have other orders. This won't be for another year - which in Bitcoin terms is like 25 years from now.
|
|
|
|
dyask
|
|
July 18, 2014, 09:33:28 AM |
|
Moore's law is approaching it's end. We won't keep seeing leaps in transistor reduction. Switching to graphene or some other exotic manufacturing might be the future.
But for what seems an eternity in the Bitcoin world, the next 1-2 years will still see more efficient miners released.
Agreed. Seriously doubt asic miniaturization will slow down within the next 18-24 months, then after that I suspect we will see a slew of optimization techniques that carry this trend on for another 6-12 months easy. Of course, all of this is largely contingent upon the bitcoin price continuing it's upward trend. I believe ALToids was saying the opposite of what you're saying. CPUs have pretty much reached a wall. ASICs for Bitcoin still have an efficient jump to 20nm, but 14nm will be cost prohibitive because all the foundries will have other orders. This won't be for another year - which in Bitcoin terms is like 25 years from now. Actually the jump to 14nm is more that just a cost problem. A lot of redesign has to be done to gain the full benefits. Probably much more than a year yet.
|
|
|
|
Golph
|
|
July 18, 2014, 03:29:55 PM |
|
http://igg.me/p/822701/xWhat do you guys think? Legit or scam? If this is real it could mean that difficulty would go up even more and our old miners would become obselete even faster. Scam, avoid at all cost
|
|
|
|
DrG
Legendary
Offline
Activity: 2086
Merit: 1035
|
|
July 19, 2014, 04:45:10 AM |
|
Moore's law is approaching it's end. We won't keep seeing leaps in transistor reduction. Switching to graphene or some other exotic manufacturing might be the future.
But for what seems an eternity in the Bitcoin world, the next 1-2 years will still see more efficient miners released.
Agreed. Seriously doubt asic miniaturization will slow down within the next 18-24 months, then after that I suspect we will see a slew of optimization techniques that carry this trend on for another 6-12 months easy. Of course, all of this is largely contingent upon the bitcoin price continuing it's upward trend. I believe ALToids was saying the opposite of what you're saying. CPUs have pretty much reached a wall. ASICs for Bitcoin still have an efficient jump to 20nm, but 14nm will be cost prohibitive because all the foundries will have other orders. This won't be for another year - which in Bitcoin terms is like 25 years from now. Actually the jump to 14nm is more that just a cost problem. A lot of redesign has to be done to gain the full benefits. Probably much more than a year yet. When producing a batch run of 100k chips, the 500K spent on 5 full time engineers is nothing compared to the $20 million needed for the run. You are right that to optimize 14nm a LOT of work needs to be done, but most ASIC companies aren't really optimizing their current 20/28nm offerings. That's why Bitfury was able to do so well on the efficiency end even with a higher node.
|
|
|
|
|