Okay let me rephrase my question, is it possible for the ASIC hardware to just do SHA256(x) instead of SHA256(SHA256(x))?

That doesn't tell me if the ASIC hardware does SHA256(SHA256(x)) or if it is SHA256(x) twice.

Does anybody know if they are explicitly double SHA256, or if they simply run the SHA256 twice.
This is important because it also denotes their usefulness outside of Bitcoin.

If that's true, then GPU mining without free electricity will be dead by the end of Winter 2014 (to account for those who try to heat their houses with GPU miners).

If an attacker was able to gain 50% hash power, they could quite possibly get upwards of an entire cycle's worth of coins.
First they put that 50% hash power into the network gradually raising the difficulty (and mining at the same time), when it adjusts the difficulty, they pull their power.
They wait the 4 weeks, then when the difficulty which is now probably 25% slower, they start mining again, and they have 25% advantage still. Which means they net a bunch coins again.
The trick would be achieving this attack for under half a million USD which would be about what they got in coins after doing this.

It seems like every time I check blockchain.info there are groups of blocks found within 30 minute blocks, and then there are times when it takes 30 minutes for a block to be found. While it still averages (according to blockchain.info) to be under 10 minutes per block, it certainly isn't that consistent.

And will this variance just get worse as the difficulty rises?

It would be interesting if somebody did this to all non BTCGuild mining pools just to push their effective hashing power to over 51% to cause trouble. But this could be very easily fixed by forcing the miners to use system time.

Don't think the government would allow that. That would give Intel a monopoly on desktop processors.

Which is 34 years before the reward is due to be zero. This is going to be a problem eventually, so why not fix it now?

Why would Satoshi (group or indiv.) use 32 bit timestamps which would overflow before the last coin was designed to be minted?

The fix is simple, just up it to 64bit unsigned integers to represent the date.

Why wouldn't they? They stand to not only increase the hashrate, but effectively ensure their product will be bought at the same time.
For every blade they sell, another one will be sold to keep somebody else in the mining game.

It is my understanding that the current generation of ASICs all get around 100MH/joule give or take a few million hashes. While that is upwards of 20x more efficient than GPU mining, how much more can be pushed out? Given we understand exactly the operation needed, does anybody know what the minimum amount of energy that could possibly be expended to perform SHA256?

If only it worked that way, we'd all be rich

It says you'll get a revenue of .02$ per day, you have to factor in your power costs.

You'll lose money overall.

Are you sure? By my calculation you'd need just over GH/J at that difficulty and price assuming .15$ per kwh. 5GH at 30W is 166MH/J

Yeah, hidden deep within my text is the subtle idea that GPU mining is dead, due to it being less efficient.

I've done some math, and come up with a formula to determine the absolute minimum sustainable (break even) network efficiency (in terms of Hashes per joule).


This basically gives the minimum efficiency that would be required to run the network if it was a single person mining and they got all the coins. So for mining to be profitable you'll need to be a tad higher than the given efficiency(to account for pool fees and luck), but it does shed light on the future of mining.

For price = .15$, difficulty as 10076292.88, reward at 3600 coins per day, and exchange rate at 110.
655714 Hashes per joule.
Which is a little less than half of what blockchain.info estimates the efficiency to be at (1.5MH/J).

Now if we fast forward  to the end of the year, I personally believe we will see 50M difficulty (350TH/s+), I also believe we will see a stable 200$ per coin, so at .15$ per kwh results in about 1.79MH/J, which cuts out some GPU miners already.
.10$ per kwh: 1.19MH/J
.05$ per kwh: 596KH/J
Obviously, cheaper electricity lowers the efficiency needed.

But what this formula really lets us know is the long term viability of the network itself. Right now, to currently survive off transaction fees alone (144*(500*.0005)) my estimate is that on average you get 500 paying transactions per block, and at the default fee, it amounts to 36 btc per day. At the current rates...
65.5MH/J are required to break even.

This makes it apparent that cheap electricity will be required to
Just some food for thought.
Now does anybody have any thoughts on this?

EDIT: I really hope my math is right, or I'll sound really stupid.

I thought the general consensus on blockchain size management was to prune spent outputs from the earlier blocks, so that the blockchain's size would be related to transaction volume instead of a history.

It's fine and dandy to believe that, except by outing this, you've put other people's information at risk. Let's say somebody does get into BFL's systems, what kind of information do you think they have stored on their servers? Information that somebody who has a vendetta against bitcoin could put to good use, such as the mailing addresses of tens of thousands of people. Not to mention any related payment information.

While, it's good of you to alert people. I think you should have alerted them instead of publicly outing their exploits as soon as you found them.

But hey, if your tactic is to get professional penetrators to cause a stir, more power to ya. I just wouldn't have done it this way.