can there be an early difficulty adjustment?

[Desolator]

When ASICs ship, if basically everyone fires them up at the same time and the total GH/s goes approx 1.5-2.0x higher, whatever difficulty it's at would result in a block around every 3 mins or something.  Is there a protection in the protocol that adjusts the difficulty early if something like that happens or do we just get to the difficulty adjustment and lose days and days off the life of bitcoin and artificially inflate the supply, both very bad things.

[jasinlee]

I dont think it will be held back at all, I think it will double from the burn ins and employees keeping their personal unit running at home.

http://betsofbitco.in/item?id=704

[FreeMoney]

When ASICs ship, if basically everyone fires them up at the same time and the total GH/s goes approx 1.5-2.0x higher, whatever difficulty it's at would result in a block around every 3 mins or something.  Is there a protection in the protocol that adjusts the difficulty early if something like that happens or do we just get to the difficulty adjustment and lose days and days off the life of bitcoin and artificially inflate the supply, both very bad things.

Huh?

and 'artificial'?

[Stephen Gornick]

Is there a protection in the protocol that adjusts the difficulty early if something like that happens or do we just get to the difficulty adjustment and lose days and days off the life of bitcoin and artificially inflate the supply, both very bad things.

The difficulty adjusts every 2,016 blocks.  If that happens in 5 days instead of 14, then it happens in 5 days instead of 14.  No big whoop.  And the target is back to 7,200 BTC per day.

Now there is a max increase though of 4X the prior difficulty.   So difficulty can't increase more than four times the previous level from one difficulty adjustment period to the next.

As far as "inflating the supply",   those were coins that were going to be entering anyway, they just were pulled forward by a few days.  An extra 5K or 10K BTC per day issued doesn't mean an extra 5K or 10K BTC hitting the exchanges, especially since unlike GPUs, most ASICs were paid well in advance and thus miner's may not be anxious to cash out.  Especially with the block reward subsidy drop to 25 coming (or already there by the time ASICs ship.)

[crazyates]

The difficulty adjusts every 2,016 blocks.  If that happens in 5 days instead of 14, then it happens in 5 days instead of 14.  No big whoop.  And the target is back to 7,200 BTC per day.

This. The diff gets calculated every 2016 blocks, so that the next 2016 blocks take ~ 2 weeks (10 min / block). If it takes only 7 days to get 2016 blocks, then the diff will double.

[Desolator]

You all do realize that if mining hardware kept getting more widespread and more powerful at an insane rate like 4x per block cycle, bitcoins would be depleted in 25 years instead of 100, right?  Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.  That's why the protocol should have been designed to overcompensate for overmining in the next cycle.  Like if 1.2x the bitcoins are mined because of an increase in GH/s, adapt the next cycle to anticipate at 0.8 total bitcoin mining rate over the next difficulty period.

[wabber]

You all do realize that if mining hardware kept getting more widespread and more powerful at an insane rate like 4x per block cycle, bitcoins would be depleted in 25 years instead of 100, right?  Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.  That's why the protocol should have been designed to overcompensate for overmining in the next cycle.  Like if 1.2x the bitcoins are mined because of an increase in GH/s, adapt the next cycle to anticipate at 0.8 total bitcoin mining rate over the next difficulty period.

Yea you are right. But if you look at the history of bitcoin it's unlikely to see a 4x of the difficulty for a long time in a row.

[vitruvio]

When the base hash rate is higher is more difficult to double it, even agreeing the  moore's law. But here don't talk about technology advances, we know about usb3 some time ago, but how many time to be used commonly?.

To reach this 4x factor inside 2 weeks that means for example that the miners will stop mining 33% of his GPU's at 350Mh/s and change to a jalapeño with 3.5 Gh/s. Not all GPU's mine at 350 MH, but to make round numbers, can use it. (substract the less power GPU's and add the FPGA).

 At actual network hash rate 21.47 Th/s (21,470 Ghs) that means that there are 21740/0.35 = 62114 devices at a mean of 350 MH/s mining right now.

To make an increase of 4x of diffifulty, need to change 33% of 62114 device = 20497, how many time will take BFL, BTFPGA.com and others to produce and sell 20,000 asics? .

In my estimation I assume some circumstances:

-when asics start to delivery no one will buy a GPU to mine, the rythm of hash rate will be setted by the rythm of asic production.
-only with an increase of 2x of difficulty a FPGA isn't profitable.
-People will change from GPU to same cost device (jalapeno) or a rig with 4 GPU's to a Single SC, bASIC or same proportion device.

So as other people says, this is not the end is only a new status quo, only asics will mine.

Correct me, but this is not the apocalypse

Regards

[vitruvio]

You all do realize that if mining hardware kept getting more widespread and more powerful at an insane rate like 4x per block cycle, bitcoins would be depleted in 25 years instead of 100, right?  Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.  That's why the protocol should have been designed to overcompensate for overmining in the next cycle.  Like if 1.2x the bitcoins are mined because of an increase in GH/s, adapt the next cycle to anticipate at 0.8 total bitcoin mining rate over the next difficulty period.

To reduce from 100 to 25 year means that you 4x the difficulty every 3.5 days over and over again, not only 1 time. If you 4x the hash power in a day (imposible even in 1 month), than means that you get 2016 blocks in a couple of days, and then the new diff make to take 2 weeks for next 2016 blocks, so bitcoins will depleted in 100 year less 12 days.

For bitcoin working (not for miners that only want to earn coins) is worse to decrease hash power than increase, if for any cause the hash power decrease to 1/4 suddenly that means that until next readjust of diff will take 2 months, the time to confirm a transaction will take 4x times, and this one of the worse characteristics of bitcoin compared with litecoin for example.

Take a look to my previous post to make some math about possibilities of difficulty increase.

Regards

[Vandroiy]

You all do realize that if mining hardware kept getting more widespread and more powerful at an insane rate like 4x per block cycle, bitcoins would be depleted in 25 years instead of 100, right?  Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.  That's why the protocol should have been designed to overcompensate for overmining in the next cycle.  Like if 1.2x the bitcoins are mined because of an increase in GH/s, adapt the next cycle to anticipate at 0.8 total bitcoin mining rate over the next difficulty period.

Let us now for a short moment consider this post, and think that difficulty indeed quadruples four times every two weeks, and thus twice a week.

Per year, that's a factor 4^(52*2). For 25 years, another exponent of 25 comes in, yielding a total network growth factor of:



That is ten to the 1565. If people were to manage a Terahash/s per atom used, the new network still vastly exceeds the size of the known universe.

"The greatest shortcoming of the human race is our inability to understand the exponential function" -Albert A. Bartlett

[jasinlee]

You all do realize that if mining hardware kept getting more widespread and more powerful at an insane rate like 4x per block cycle, bitcoins would be depleted in 25 years instead of 100, right?  Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.  That's why the protocol should have been designed to overcompensate for overmining in the next cycle.  Like if 1.2x the bitcoins are mined because of an increase in GH/s, adapt the next cycle to anticipate at 0.8 total bitcoin mining rate over the next difficulty period.

Let us now for a short moment consider this post, and think that difficulty indeed quadruples four times every two weeks, and thus twice a week.

Per year, that's a factor 4^(52*2). For 25 years, another exponent of 25 comes in, yielding a total network growth factor of:



That is ten to the 1565. If people manage to deliver a Terahash from one atom, the new network still vastly exceeds the size of the known universe.

"The greatest shortcoming of the human race is our inability to understand the exponential function" -Albert A. Bartlett

+1 awesome.

[vitruvio]

You all do realize that if mining hardware kept getting more widespread and more powerful at an insane rate like 4x per block cycle, bitcoins would be depleted in 25 years instead of 100, right?  Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.  That's why the protocol should have been designed to overcompensate for overmining in the next cycle.  Like if 1.2x the bitcoins are mined because of an increase in GH/s, adapt the next cycle to anticipate at 0.8 total bitcoin mining rate over the next difficulty period.

Let us now for a short moment consider this post, and think that difficulty indeed quadruples four times every two weeks, and thus twice a week.

Per year, that's a factor 4^(52*2). For 25 years, another exponent of 25 comes in, yielding a total network growth factor of:



That is ten to the 1565. If people were to manage a Terahash/s per atom used, the new network still vastly exceeds the size of the known universe.

"The greatest shortcoming of the human race is our inability to understand the exponential function" -Albert A. Bartlett

I think I call this with a only word, impossible.

[DannyHamilton]

You all do realize that if mining hardware kept getting more widespread and more powerful at an insane rate like 4x per block cycle. . .

You do understand that the math you present is not physically possible, right?

Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.

Usable life?  I don't understand.  How does mining coins faster cause a reduction in useable life?

[os2sam]

Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.

Usable life?  I don't understand.  How does mining coins faster cause a reduction in useable life?

I was going to comment on that as well.  It sounds like he thinks that when all coins are mined bitcoin will cease to exist/be usable???  Bitcoin will live on as long as there is a network and hashs are being done.
Sam

[Desolator]

well obviously when coins become unmineable, a couple things will happen:

1. bitcoins will continue to slowly get "lost" from improperly backed up wallets, etc, until there is an insufficient amount to go around.  That will take A LONG time but still.
2. the price will either go through the roof from continuing demand or will crash as people lose faith in the current system for one of many reasons or quit because a similar fork has mining for $$$ still available
3. everyone will turn their miners off unless pools play nice and share the transaction fees with miners (over 50% of them right now do not) which would open up the system to a >50% attack.

and more bad things that will happen over time regardless that are made worse by a massive acceleration are:
1. deflation due to way more coins being generated but the same demand from non-miners
2. the block chain database gets huuuuuge and it's even harder for bitcoin clients to catch up.  I left my client closed for about 3 months on one of my PCs and 72 hours later of running it on and off, it's still 5500 blocks out of date and this is an i5 with 8GB of RAM and a 1TB high speed Seagate hard drive on a 10 megabit connection.  What do new people do? Run it for 2 weeks? lol.

[DannyHamilton]

1. bitcoins will continue to slowly get "lost" from improperly backed up wallets, etc, until there is an insufficient amount to go around.  That will take A LONG time but still.

When you say "A LONG time", how long are you talking about.  You are saying a few thousand years, right?  I'm not sure that I need to be worrying about what the state of the Earth's economy will be like in a few thousand years.

2. the price will either go through the roof from continuing demand or will crash as people lose faith in the current system for one of many reasons or quit because a similar fork has mining for $$$ still available

But this is true every single day, right?  On any day (even today) the price could go through the roof from increased demand or could crash if people suddenly lose faith in the current system. A similar fork could even be created today that has continuous mining built in, right?  I don't see how any of this requires waiting for bitcoins to become unminable.
 

3. everyone will turn their miners off unless pools play nice and share the transaction fees with miners (over 50% of them right now do not) which would open up the system to a >50% attack.

It should be rather obvious that as the quantity of newly minted coins goes down, people will shift to pools that share some of the transaction fees.  To remain functional the other pools will be forced to compete by sharing some of their transaction fees as well.  If some people turn off their miners, this will reduce the difficulty and increase the profitability of mining for those who continue to mine.  Eventually an equilibrium will be reached where enough miners find it profitable enough to keep their miners running.

more bad things that will happen over time. . .
1. deflation due to way more coins being generated but the same demand from non-miners

Bitcoin is intentionally designed to start out inflationary and then slowly shift to deflationary.  This is not a "bad thing".  This is intentional in the design.

2. the block chain database gets huuuuuge and it's even harder for bitcoin clients to catch up.

It is expected that blockchain pruning and light clients will address this issue eventually.

[os2sam]

well obviously when coins become unmineable, a couple things will happen:

1. bitcoins will continue to slowly get "lost" from improperly backed up wallets, etc, until there is an insufficient amount to go around.  That will take A LONG time but still.
2. the price will either go through the roof from continuing demand or will crash as people lose faith in the current system for one of many reasons or quit because a similar fork has mining for $$$ still available
3. everyone will turn their miners off unless pools play nice and share the transaction fees with miners (over 50% of them right now do not) which would open up the system to a >50% attack.

and more bad things that will happen over time regardless that are made worse by a massive acceleration are:
1. deflation due to way more coins being generated but the same demand from non-miners
2. the block chain database gets huuuuuge and it's even harder for bitcoin clients to catch up.  I left my client closed for about 3 months on one of my PCs and 72 hours later of running it on and off, it's still 5500 blocks out of date and this is an i5 with 8GB of RAM and a 1TB high speed Seagate hard drive on a 10 megabit connection.  What do new people do? Run it for 2 weeks? lol.

Have you ever read the bitcoin white paper?  It answers many those concerns.

If/when Bitcoins are lost they will be denominated into more decimal places.  There is no reason to stop at a Satoshi (8 decimal places).

If Bitcoin is adopted it will dramatically increase in value to the point of the currency of the entire world being 2 Million Bitcoin's minus the lost coins.

As more mine reward halving happens the transaction fees will be the sole source of income for mining and if pools are still around then they will have to distribute transactions fees to miners.

As for the large block chain the new clients are supposed to take that into account.  What client are you using?  Maybe you need to upgrade it.
Sam

[Aseras]

You all do realize that if mining hardware kept getting more widespread and more powerful at an insane rate like 4x per block cycle, bitcoins would be depleted in 25 years instead of 100, right?  Cutting off like 7 days may not seem bad but any reduction to the useable life is bad.  That's why the protocol should have been designed to overcompensate for overmining in the next cycle.  Like if 1.2x the bitcoins are mined because of an increase in GH/s, adapt the next cycle to anticipate at 0.8 total bitcoin mining rate over the next difficulty period.

It'll end up like namecoid did a while back with months between difficulty changes.