ASICMINER: Entering the Future of ASIC Mining by Inventing It

[jimmothy]

Everyone else understands that 20nm, if ever they get it working and yields become reasonable, should provide lower production cost per GH as well as better power efficiency

I'm dense? Your repeating the same useless argument over and over that they "could" increase efficiency but haven't.

If they can't acheive less than 0.5w/gh while still below $0.2/gh wafer cost then it is meaningless. And I doubt they can or they would have already.

And does "everyone" include nvidia? They seem to think 20nm wafer costs are too expensive.

simply by lowering clocks and voltages from near the top of the schmoo plot to somewhere lower. The same will work for your GPU, for your CPU (both of which will in fact do this automatically when mostly idle)

You seem to think lowering voltage can guarantee higher efficiency. Show me some proof. Has it ever occurred to you that chip manufacturers may have already set voltage as low as possible without diminishing returns?

You repeat over and over that it has to work because "physics" and yet in reality there is not a single person with an asic who acheived less than 0.5w/gh at the chip level.

Why? Laziness?

[1714955432]

1714955432

[1714955432]

1714955432

[1714955432]

1714955432

[1714955432]

1714955432

[raskul]

Everyone else understands that 20nm, if ever they get it working and yields become reasonable, should provide lower production cost per GH as well as better power efficiency

I'm dense? Your repeating the same useless argument over and over that they "could" increase efficiency but haven't.

If they can't acheive less than 0.5w/gh while still below $0.2/gh wafer cost then it is meaningless. And I doubt they can or they would have already.

And does "everyone" include nvidia? They seem to think 20nm wafer costs are too expensive.

simply by lowering clocks and voltages from near the top of the schmoo plot to somewhere lower. The same will work for your GPU, for your CPU (both of which will in fact do this automatically when mostly idle)

You seem to think lowering voltage can guarantee higher efficiency. Show me some proof. Has it ever occurred to you that chip manufacturers may have already set voltage as low as possible without diminishing returns?

You repeat over and over that it has to work because "physics" and yet in reality there is not a single person with an asic who acheived less than 0.5w/gh at the chip level.

Why? Laziness?

dude, you really could cause an argument in an empty room. chill down.

[AMuppInTime]

Jimmothy is a troll, just put him on ignore. Please don't quote him because it forces the rest of us to read him.

[lophie]

Jimmothy is a troll, just put him on ignore. Please don't quote him because it forces the rest of us to read him.

I thought he is a "sloth" not a "troll"

[jimmothy]

dude, you really could cause an argument in an empty room. chill down.

You do realize I was defending your favorite company which you speak about at every chance right?

Am I wrong to think that spondoolies next gen will be in an entirely seperate class of asics compared to what is currently on the market?

Without any testing or a single shred of evidence do you really think current gen asics can be simply undervolted to reach below 0.38w/gh while still being cost effective?

[Jutarul]

dude, you really could cause an argument in an empty room. chill down.

You do realize I was defending your favorite company which you speak about at every chance right?

Am I wrong to think that spondoolies next gen will be in an entirely seperate class of asics compared to what is currently on the market?

Without any testing or a single shred of evidence do you really think current gen asics can be simply undervolted to reach below 0.38w/gh while still being cost effective?

Without going into the debate with competitors too much - from what is observable from the progression in bitcoin mining IT is that gen3 equivalent hardware from AM and "next-gen" hardware will likely co-exist long-term (> 2 years). This is because there are 10-fold variations in energy prices dependent on location and still large bitcoin price fluctuations, which gives mining a timing value. As such one may expect the less efficient hardware to migrate into areas of low electricity costs and opportunistic mining, while high-efficiency hardware goes into high-cost areas (e.g. through supply and maintenance chain issues) and sustained mining. This will make the overall hashing power of the network much more elastic.

The bottom line is that while gen1 or gen2 ASICs may come to an abrupt death, the gen3 ASICs will likely fade out smoothly.

[jimmothy]

Without going into the debate with competitors too much - from what is observable from the progression in bitcoin mining IT is that gen3 equivalent hardware from AM and "next-gen" hardware will likely co-exist long-term (> 2 years). This is because there are 10-fold variations in energy prices dependent on location and still large bitcoin price fluctuations, which gives mining a timing value. As such one may expect the less efficient hardware to migrate into areas of low electricity costs and opportunistic mining, while high-efficiency hardware goes into high-cost areas (e.g. through supply and maintenance chain issues) and sustained mining. This will make the overall hashing power of the network much more elastic.

The bottom line is that while gen1 or gen2 ASICs may come to an abrupt death, the gen3 ASICs will likely fade out smoothly.

I agree with your end game prediction about hardware distribution and that gen3 will fade out much slower than previous generations but I don't think it will take 2+ years. The market will be flooded with 14nm asics by then.

Although there is a 10-fold variation in electricity prices, mining will only be feasible for the top few percent. It really doesn't make much sense to mine with $0.2+/kwh.

[Jutarul]

Although there is a 10-fold variation in electricity prices, mining will only be feasible for the top few percent.

I never understood the tendency to view mining cost projection in terms of a static equilibrium. It's dynamic and electricity prices are just one factor. Please enlighten me if you think mining doesn't fall into the realm of steady state systems (http://en.wikipedia.org/wiki/Steady_state).

It really doesn't make much sense to mine with $0.2+/kwh.

Of course it makes sense - at least for the latest and more efficient equipment. You have to compare the opportunity cost of shipping and maintaining hardware at a remote location with the opportunity to keep it local. Relocation is a major effort and the timing is essential for the lifetime yield of new hardware.

[Swordsoffreedom]

Jimmothy is a troll, just put him on ignore. Please don't quote him because it forces the rest of us to read him.

I thought he is a "sloth" not a "troll"

^_^ That made me smile a bit well played Lophie it is a sloth.
As for the technical discussion mining is dynamic since it follows that increased processing power is constantly required to mine.
Factoring in energy costs to determine if something is profitable always needs to be taken into account.
Example time it takes for chips to hit mass scale can seriously be impact profits with a month or two in this space.
That said FC should be fine based on the timescales we know.

[minerpumpkin]

I find the thought of jimmothy being a sloth highly amusing. Imagine a sloth, literally the animal, sitting in front of a computer researching ASIC specifications and investing in ASICMiner.
And I, personally, like his optimism!

[lophie]

If you liked my sloth comment get me a cup of coffee or beer

(1LophieEaKWKtqGWoTDp5TDhB7rGjneHep),

make my day 

[aahzmundus]

Without going into the debate with competitors too much - from what is observable from the progression in bitcoin mining IT is that gen3 equivalent hardware from AM and "next-gen" hardware will likely co-exist long-term (> 2 years). This is because there are 10-fold variations in energy prices dependent on location and still large bitcoin price fluctuations, which gives mining a timing value. As such one may expect the less efficient hardware to migrate into areas of low electricity costs and opportunistic mining, while high-efficiency hardware goes into high-cost areas (e.g. through supply and maintenance chain issues) and sustained mining. This will make the overall hashing power of the network much more elastic.

The bottom line is that while gen1 or gen2 ASICs may come to an abrupt death, the gen3 ASICs will likely fade out smoothly.

I have been saying this for a while.  Someone recently pointed out that in some parts of the developing world, in remote as fuck areas, the government subsidies electricity to the point where it is far cheaper then it should be in an effort to attract industry.  There are very few takers as who wants to have a factory stuck in the middle of Africa.  But bitcoin mining....

So who wants to do the leg work and get an ASIC farm set up in one of these areas!

[rudi]

Also, in some places in the world (e.g. France), heating is done with electricity.
Since heating is done with electricity anyway, it would seem to make sense to swap the electrical resistance with cheap bitcoin mining ASICs.

[atomicchaos]

I have been saying this for a while.  Someone recently pointed out that in some parts of the developing world, in remote as fuck areas, the government subsidies electricity to the point where it is far cheaper then it should be in an effort to attract industry.  There are very few takers as who wants to have a factory stuck in the middle of Africa.  But bitcoin mining....

So who wants to do the leg work and get an ASIC farm set up in one of these areas!

Uh, it's not just remote areas.. I'm paying .05 kw/h (no delivery cost) in the US. Of course the place that has this cheap power is a very small town with the town running things. Maybe this isn't as cheap as you're referring to, but it has allowed me profitability when most are unprofitable.

Was thinking of opening a second site that I found available as it's cheaper than upgrading my 400 amp 3 phase service to go to another location with existing available service. This majority of the town also uses electricity to heat their houses in the winter.

[MrWDunne]

Without going into the debate with competitors too much - from what is observable from the progression in bitcoin mining IT is that gen3 equivalent hardware from AM and "next-gen" hardware will likely co-exist long-term (> 2 years). This is because there are 10-fold variations in energy prices dependent on location and still large bitcoin price fluctuations, which gives mining a timing value. As such one may expect the less efficient hardware to migrate into areas of low electricity costs and opportunistic mining, while high-efficiency hardware goes into high-cost areas (e.g. through supply and maintenance chain issues) and sustained mining. This will make the overall hashing power of the network much more elastic.

The bottom line is that while gen1 or gen2 ASICs may come to an abrupt death, the gen3 ASICs will likely fade out smoothly.

I have been saying this for a while.  Someone recently pointed out that in some parts of the developing world, in remote as fuck areas, the government subsidies electricity to the point where it is far cheaper then it should be in an effort to attract industry.  There are very few takers as who wants to have a factory stuck in the middle of Africa.  But bitcoin mining....

So who wants to do the leg work and get an ASIC farm set up in one of these areas!

You seem to have neglected the fact that you would have to cool these machines. Thats a lot easier to do in a colder country.

[punin]

You seem to think lowering voltage can guarantee higher efficiency. Show me some proof. Has it ever occurred to you that chip manufacturers may have already set voltage as low as possible without diminishing returns?

You repeat over and over that it has to work because "physics" and yet in reality there is not a single person with an asic who acheived less than 0.5w/gh at the chip level.

Here's some proof, oh yeah and BTW our chip on 55nm achieved that "less than 0.5W/Ghash" back in june 2013

[eyebrows]

Will their be usb sticks 4sale, too?

[Puppet]

Here's some proof, oh yeah and BTW our chip on 55nm achieved that "less than 0.5W/Ghash" back in june 2013

Seems like you even went below 0.4W/GH at the chip level. Not bad for your first ever asic design (IIRC) on an obsolete process and nearly one year ahead of the competition. Cant wait to see what you will do on a more advanced process. Havent kept up, is something upcoming ?

That said, wouldnt waste my time trying to talk sense to "jimmothy". Logic is lost on him.

[jimmothy]

You seem to think lowering voltage can guarantee higher efficiency. Show me some proof. Has it ever occurred to you that chip manufacturers may have already set voltage as low as possible without diminishing returns?

You repeat over and over that it has to work because "physics" and yet in reality there is not a single person with an asic who acheived less than 0.5w/gh at the chip level.

Here's some proof, oh yeah and BTW our chip on 55nm achieved that "less than 0.5W/Ghash" back in june 2013

Well that's about as solid proof as it gets. Thanks for proving me wrong without theoretical evidence.

0.38w/gh is incredibly impressive but I want to know is it still cost effective while only 1gh per chip?

And puppet this in no way validates your claims that any and all asic manufacturers can increase efficiency beyond what was advertised. Bitfury is a rare exception and even then 0.38w/gh was advertised I was just unaware of it.

Please note that it seems that 96 Mhz is close to _BEST_ solution

This leads me to beleive that even bitfury is towards the deep end of the shmoo plot.

How do we know other manufacturers are not?

And even if all manufacturers can acheive below 0.38w/gh can they do it cost effectively?

[klondike_bar]

You seem to think lowering voltage can guarantee higher efficiency. Show me some proof. Has it ever occurred to you that chip manufacturers may have already set voltage as low as possible without diminishing returns?

You repeat over and over that it has to work because "physics" and yet in reality there is not a single person with an asic who acheived less than 0.5w/gh at the chip level.

Here's some proof, oh yeah and BTW our chip on 55nm achieved that "less than 0.5W/Ghash" back in june 2013

Well that's about as solid proof as it gets. Thanks for proving me wrong without theoretical evidence.

0.38w/gh is incredibly impressive but I want to know is it still cost effective while only 1gh per chip?

And puppet this in no way validates your claims that any and all asic manufacturers can increase efficiency beyond what was advertised. Bitfury is a rare exception and even then 0.38w/gh was advertised I was just unaware of it.

Please note that it seems that 96 Mhz is close to _BEST_ solution

This leads me to beleive that even bitfury is towards the deep end of the shmoo plot.

How do we know other manufacturers are not?

And even if all manufacturers can acheive below 0.38w/gh can they do it cost effectively?

look at  bitmain - the S1 is ~2w/GH which the S2 is 1w/GH, by running each chip at about 60% the power.  Presumably less than 1w/GH could be done, but would require many more chips, which is cost-prohibitive for the time being