Bitfury 16nm ASIC Demo - Part 3 of 3 here, immersion cooling.
https://www.youtube.com/watch?v=DSbmuBvDWo0
Topic: Video: Bitfury 16nm ASIC Demo - Part 1 of 3 (Read 1740 times)
Great efficiency increase 
Less power draw the better. Gonna have to wait to see how they perform in the real world.
Less power draw the better. Gonna have to wait to see how they perform in the real world. 
Thanks for links to video 1 and 2.
I would love to see a longer test.
But they look really nice.
Good luck with sales of chips.
I would love to see a longer test.
But they look really nice.
Good luck with sales of chips.
Part 2 of 3 https://www.youtube.com/watch?v=zZPum2zqGPE
I really hope we get some USB miners made with these chips!
Ja. If they themselves are not interested in getting into the 'selling shovels to the (gold) miners' business then most definitely they should either just non-exclusively partner with an assembler/seller of miners or release a reference design and sell chips to whomever wants whatever MOQ that is enough for them to bother with.
I'm just glad that Bitfury is doing a proper announcement of these chips, making it known that they have a viable chip and it is in testing. Period. No taking pre-orders from the general public on what might or might not work as opposed to taking the BFL (Monarch)/Bitmine.ch (A1) route of grabbing all the cash they can before even having engineering samples.
I'd think that it is a safe bet Bitmain has something similar in the works but being Bitmain they are quiet about it so current chip/miner sales are not impacted.
As for reverse engineering... The chips? Decapping them to see the innards of it would be of no help to Bitmain/Avalon/whoever. While block and signal path layout is critical to max performance (eg Bitmain's 'optimized' chips in the s7) , seeing what Bitfury did is a waste of time. Chip designers know how to do it right, the problem usually is that project managers rarely take that route because it costs (and risks) $$$ vs using pre-packaged IP blocks provided by the chip foundries.
As has always been the case the biggest hurdles to 16/14nm gate-size chips of any sort is the process itself. Only late last year did TSMC take their 16nm chip production from limited boutique chips & engineering sample runs to full open production. Once the few foundries that are capable of running 16/14nm chips have the process nailed down enough to lower the scrap rate you will see everyone and their brother jumping on the bandwagon.
I'd think that it is a safe bet Bitmain has something similar in the works but being Bitmain they are quiet about it so current chip/miner sales are not impacted.
As for reverse engineering... The chips? Decapping them to see the innards of it would be of no help to Bitmain/Avalon/whoever. While block and signal path layout is critical to max performance (eg Bitmain's 'optimized' chips in the s7) , seeing what Bitfury did is a waste of time. Chip designers know how to do it right, the problem usually is that project managers rarely take that route because it costs (and risks) $$$ vs using pre-packaged IP blocks provided by the chip foundries.
As has always been the case the biggest hurdles to 16/14nm gate-size chips of any sort is the process itself. Only late last year did TSMC take their 16nm chip production from limited boutique chips & engineering sample runs to full open production. Once the few foundries that are capable of running 16/14nm chips have the process nailed down enough to lower the scrap rate you will see everyone and their brother jumping on the bandwagon.
even better. my chip building knowledge is very little. So if reverse engineering being a
non-issue. They really need to sell a lot of chips to a lot of builders/ dev
I'm just glad that Bitfury is doing a proper announcement of these chips, making it known that they have a viable chip and it is in testing. Period. No taking pre-orders from the general public on what might or might not work as opposed to taking the BFL (Monarch)/Bitmine.ch (A1) route of grabbing all the cash they can before even having engineering samples.
I'd think that it is a safe bet Bitmain has something similar in the works but being Bitmain they are quiet about it so current chip/miner sales are not impacted.
As for reverse engineering... The chips? Decapping them to see the innards of it would be of no help to Bitmain/Avalon/whoever. While block and signal path layout is critical to max performance (eg Bitmain's 'optimized' chips in the s7) , seeing what Bitfury did is a waste of time. Chip designers know how to do it right, the problem usually is that project managers rarely take that route because it costs (and risks) $$$ vs using pre-packaged IP blocks provided by the chip foundries.
As has always been the case the biggest hurdles to 16/14nm gate-size chips of any sort is the process itself. Only late last year did TSMC take their 16nm chip production from limited boutique chips & engineering sample runs to full open production. Once the few foundries that are capable of running 16/14nm chips have the process nailed down enough to lower the scrap rate you will see everyone and their brother jumping on the bandwagon.
I'd think that it is a safe bet Bitmain has something similar in the works but being Bitmain they are quiet about it so current chip/miner sales are not impacted.
As for reverse engineering... The chips? Decapping them to see the innards of it would be of no help to Bitmain/Avalon/whoever. While block and signal path layout is critical to max performance (eg Bitmain's 'optimized' chips in the s7) , seeing what Bitfury did is a waste of time. Chip designers know how to do it right, the problem usually is that project managers rarely take that route because it costs (and risks) $$$ vs using pre-packaged IP blocks provided by the chip foundries.
As has always been the case the biggest hurdles to 16/14nm gate-size chips of any sort is the process itself. Only late last year did TSMC take their 16nm chip production from limited boutique chips & engineering sample runs to full open production. Once the few foundries that are capable of running 16/14nm chips have the process nailed down enough to lower the scrap rate you will see everyone and their brother jumping on the bandwagon.
50gh a chip at 3.5 watts would be the best usb sticks ever made.
I always I am willing to demo any made product. It would really be smart to get this to market in march and flood the market with the chips. Scatter them all over.
We all know that people will reverse engineer them.
So a really big release of chips would be nice.
I would love to have some .06 to .10 watt gear.
Very cool. Imagine 20TH at 1200 watts. 
it would only cause a huge diff increase, so aside from the initial earning, for a shor time frame nothing will really change
but al those miners with great efficiency, usually come first in china, this mean that those mienrs are actually promoting centralization
If distributed correctly the might end with the domination of China in the mining race...
But I don´t think it can happen....
Interesting development. Nice efficiency increase.
Very cool. Imagine 20TH at 1200 watts.
it would only cause a huge diff increase, so aside from the initial earning, for a shor time frame nothing will really change
but al those miners with great efficiency, usually come first in china, this mean that those mienrs are actually promoting centralization
Impressive chip results! I can't wait to see footage of the 2PIC data center where they plan to host these things.
“We are planning to roll out the products starting from late Q1 2016. Price point will be determined by the total performance of the product and will be lower by $/GH than anything that is available on the market at this time.”
We will see..........
We will see..........
This next gen of hardware may change things for the average miner with low overhead.
Wow, this is crazy good, hope they would send in random to some to test it for their personal mining experience, can't wait to mine using one of these. Couldn't find a rumour about how much this is going to be around, anyone knows?
Very cool. Imagine 20TH at 1200 watts.
it would only cause a huge diff increase, so aside from the initial earning, for a shor time frame nothing will really change
but al those miners with great efficiency, usually come first in china, this mean that those mienrs are actually promoting centralization
Very cool. Imagine 20TH at 1200 watts.
Jump to: