You have a very limited number of stream processors at your disposal.  All you can do is try to overclock.  Don't try to use vectors or any fancy settings in your miner.

I give this thread 2.3 out of 10.0.  If you want to learn how to whine properly about people setting prices according to their own interests and not yours, head over to the EVE Online forums, Market Discussions.

Worse than that.  The mainline client needs to be changed, plus all of the pools, plus all of the miners.

Excellent idea, so I'm sure it'll happen.  But it will take a little while.

Enough means, well, enough.  If you run out, it is bad and the peg is broken.

But, if you control your local currency, and are a net exporter to the US, you can run it until political pressure forces you to stop doing it.

Consider China.  They control their national currency, so they can always buy dollars and sell Renminbi, because they can create Renminbi out of thin air.  They can also sell dollars and buy Renminbi, because as a net exporter to the US, they always have more dollars coming in than they need.

For an overview of how a currency peg works in practice, see this article.  http://www.marketskeptics.com/2009/01/hyperinflation-will-begin-in-china-and.html

I'm not sure about the conclusions they draw, but the first diagram is worth a thousand words in understanding how it all fits together.

A peg works when a central bank has enough of the foreign currency that they can credibly offer to exchange to and from the local currency at a fixed ratio.  Since the US Dollar is a huge major reserve currency around the world, most foreign central banks, particularly in countries that export to us, are able to do this, if they choose to.

Why they would want to do so is a more difficult question.

Read it again.  It doesn't say anything even remotely resembling what you are claiming it says.

Generating a valid block after the rest of the network has already moved on to the next one is very nearly pointless.  Do you understand what circumstances must be present for someone to replace the existing chain with a side branch of their own?  How many times do you think they could get away with what the network operators would consider to be a very major event before someone notices?

Sorry, you are just completely wrong about this.  Validity of a transaction is not determined by being in a block or not.  A transaction that isn't in a block merely hasn't been accepted (yet).

Say you had 51% of the total mining capacity.

What are the odds of getting the next block, and being able to prevent transactions that you don't approve of for ~10 minutes?  51%
And the odds of getting the block after that and continuing the blockade out to ~20 minutes?  26%
Add if you need ~30 minutes?  13%
40 minutes?  7%
50 minutes?  3%
An hour?  2%
Two hours?  0.031%

Difficulty has nothing to do with mining odds.  Difficulty acts to slow down the global network, it does nothing to change the distribution of block finding.

The odds of you finding the next block first depend only on the fraction of global hashing power that you possess.

However, the latest few difficulty increases imply that the network has seen and expects to continue seeing roughly 3.5% daily increases in hashing power.  Now obviously we can't sustain 3.5% growth forever.  The bad news is that it really doesn't have to continue for very long at all to invalidate most other projections.

Note that Fergalish did his math at the dawn of a new era, when CPUs gave way to GPUs with hundreds of times more power.  I see no evidence of such a shift happening now.  People are simply brining new GPUs online at a rate of around 3.5% per day.

Mining is how transactions get processed.  Every 4 years or so, the fixed part of the blockfinding reward halves.  The theory is that the transaction fee will gradually increase as that happens to ensure that miners keep getting rewarded.

Current global hashing power: 1.97E+12 Hashes / Second (source)

That puts the odds of any given 1 Mhash/sec winning a block at 0.0000506%.  Over the course of "today" (the next 144 blocks) it has a 0.00729% of finding a block.

The current growth in difficulty suggests an expected growth in global hashing power at 3.5% per day.  That means that tomorrow, the chance per block will be 0.0000489%, and overall 0.00705%.

Assuming this rate of growth, the limit (sum of the series over an infinite lifetime) is 0.215% chance of finding a block.

So, a GPU, or whatever, with 300 Mhash/sec of power has about a 65% chance of finding a block between now and the end of time.  The expected payout is thus 32.34 BTC, or about $226 at $7/BTC.

Thus, with my pessimistic assumptions, you should not spend more than 75 cents per Mhash/sec, which is still possible, but getting harder.

Note also that I'm assuming zero operating expenses (free power and cooling).

The retarget process takes the rate that was appropriate a week (1008 blocks) ago and increases it by the amount necessary to make it appropriate a week (1008 blocks) from now.

The jump from the before block to the after is huge because it includes 2 weeks of growth in an instant.  In this case, about 3.5% daily growth.

If the global hashing pool had grown by a factor of 10 a day or two ago, it would have had nearly no effect on the current target.  Instead you'd have seen is a huge reduction in the mean block time, which has not happened.

No, no, no.  This is completely 100% WRONG.  You don't seem to understand how any of this works at all.

With more than 50% of the hashing power, you can hide a double spend for a few minutes, and only from funds that you have legitimately.  You can not counterfeit anything.

I still find absolutely no evidence of a massive hashing machine, whether powered by unicorn dust or by FPGAs.

One person has been making very sketchy claims, which of course he can't provide any evidence for because the Men In Black offered him 25% to keep quiet.  That is 25% of a venture that is highly unlikely to be profitable even if they didn't have to pay him hush money, by the way.

If you think that the recent jump in difficulty was caused by a mythical super-cluster coming online in the last day or two, you really need to go back and read up on the retargeting process.

We will have changed from 64 to 128 bit floats for technical reasons LONG before any need to do it for economic reasons.

Look here.  All the information you could ever dream of.

But their transactions are still valid, and you have no way of mapping a wallet to a (node authentication) account, so they can still use the coins.  And in the scenario you describe, mining is cheating.  A good cheat won't necessarily crank up and claim every block for himself.  But if he did, would you know?  Would you know who?  The only way to identify which node on the network was the origin of a valid block is if every node attaches to you.

Why not just piggyback on the global system?  They can't compete with the big network, so there will be no free money, but they would still get all of the advantages of the currency.

I'm not sure that bitcoin is a good fit for what you need.  It assumes an abundance of power, both electrical and computational.

Bitcoin makes it hard to forge blocks by making each block require as much power as the network can possibly throw at it.  You can't keep people from cheating on your extra rule (no easy access to power) by rewarding them if they do (by having enough power to overwhelm your network).

All they'd really have to do would be to set up and maintain a hashing fleet with enough power to monopolize the block generation stages, and then refuse to include transactions that don't have an appropriate fee.

About the best I can think of is that if the hashing function doesn't take 100% effort, cheating is no longer impossible.  When the hashing function is as hard as the global system can stand, it is mathematically certain that the block is legit, just as 2+2=4 is mathematically certain.

If there was some sort of useful work being done as a proxy for the hashing function, we are relying on a different, and unprovable, security scheme.

Right now, increases in hashing power directly translate into increased difficulty.  When we found a way to cheat at hashing (by using GPUs), hashing automatically became harder.  There is no reason to think that this property would hold for a proxy.