Only there is only about one valid nonce in 8,974,296 candidate blocks now. Meditate on this and you will be enlightened.

I hereby grant you exactly the salary I receive for working on Bitcoin.

Then why are you not hysterically warning about the miner who mined http://blockexplorer.com/block/00000000000000001e8d6829a8a21adc5d38d0a473b144b6765798e61f98bd1d having enough hashpower to replace the entire history of Bitcoin in an hour?

It's worth pointing out for posterity that you've gone and edited your claims in this thread to back them off from their previously falsified forms. This is somewhat discourteous to the other participants here, because it makes it look like you didn't take the positions that they've argued with.  I'm glad to see more accurate claims being made, but I'm wondering when you're going to go revise them further to point out that you were completely wrong and that your extranonce data strongly supports the position that there were multiple people mining the whole time.

Yup. Exactly. But that kind of view doesn't let you generate ALL CAPS headlines.

Rolling difficulty roughly halves the computational cost compared to quantized difficulty to mine a fork down to diff 1 in order to keep an isolated node tricked with a fake chain.

Generally these schemes are vulnerable to subtle design and implementation errors. Personally, I don't think it's worth worrying about: If you lose more than 3/4 of your hashrate you have much bigger problems— and you can recover with a manual update that just resets the difficulty to be wherever you want it.

Why did you claim it was only reset on restart?

How does this contradict anything at all that I or death and taxes said?  In fact, it seems to __thoroughly__ refute the claim you were making that Satoshi mined alone the first year. Unlike you, however, I'm not going back and editing my posts when people point out that they cannot possibly be correct.

Why do you keep posting all these claims without even bothering to look at the source?

When you say you haven't checked 'all versions' you mean to say 'any versions except one really old one'?

Currently it's reset on every block:

I don't mind correcting you because that took all of two seconds ... but ... really?  "vulnerability"?

What is your motivation in not spending two seconds to actually look before claiming that "all are affected by the vulnerability"?

Even back in October 2010 (e.g. v0.3.14):
It also spent some time with the wrap at 0x7f.

I can most likely have 1000 pcs of artix on my doorstep tomorrow morning. The same is not true for Avalon— AFAIK they've not given much in the way of dates, but it looks like their pipeline is a couple months long (though hopefully shortening).  Re-usability is a benefit. It's not the only one.

I would expect them to be substantially _less_ costly per MH.  E.g. fitting potentially 400-900MH/s in an $100 part. Beyond the basic process improvements the Artix/Kintex parts have a substantially improved architecture.  Of course, the S6LX150 designs have had a lot of effort applied to optimizing them.

A batch 3 Avalon cost 75 BTC and delivers 68GH/s. At the current market price that is  $75/GHs.   So maybe a 28nm FPGA ends up being twice the price per MH initially.   But: It would have a smaller basic unit size (probably a 2 or 4 chip board) which would create a larger customer base and it could potentially have better supply.  What does it matter that _in theory_ an Avalon is $75/GHs now when you can't actually buy one??

Plus, once there is a ASIC miner on a better process in widespread use the avalons will either need to head to the landfill— or at least Alaska for use as heating units, while a 28nm FPGA may have better reusablity.

I imagine that if someone had a board like the x6500 based on artix _right now_ they would be selling out. In a month? who knows.

So I don't think this is interesting just as a novelty project... though if it's just an "okay idea that won't make much money" vs an amazing one depends a lot on how the asic miner supply plays out, how the bitcoin market price plays out, etc.

I think it's highly questionable that Opencoin's Ripple consensus model will achieve both decentralization and security (one or the other seems more likely).  See the thread I created at https://bitcointalk.org/index.php?topic=144471.0 in particular the later posts where it show some topology diagrams.

I'd love to understand what the simple necessary and sufficient criteria is for ripple's consensus to be secure, but I don't think _anyone_ (not even the people creating ripple) actually know precisely what the criteria are.

The really high end FPGAs are targeted at people doing large scale logic simulations and other stuff that _cannot_ be done in a smaller part and so they have a premium price, they also often have very fast specialized IO hardware and other stuff that mining doesn't need.  I wouldn't expect them to be the most effective for mining. I expect mining to be most cost effective on whatever the largest or second largest low end devices is in any product line— keeping down the per chip costs and also dodging premium product pricing.

Of course, nothing beats actually simulating a design.

Hi all.   I'm really surprised that vanitygen hasn't switched to generating compressed keys.

Uncompressed keys have public keys about twice the size of compressed ones and this means that when you use vanitygen produced addresses you end up paying more in transaction fees, causing more increase in blockchain size, etc.

I have modified my own copy of vanitygen to produce compressed keys— and as a bonus it's about 15% faster. ... but I've only done CPU support, and at the moment I no longer have any ati GPUs (having sold off my GPU mining farm)... and I can't really justify putting in time to add support for features I don't need and thoroughly test it when I have so many other things I need to do for myself and others.

Are people still using vanitygen?

A few people seem to be getting thrown by the mistaken belief that Litecoin's SCRYPT is ASIC incompatible— it's true scrypt itself was designed to flatten the playing field between specialized hardware an general purpose hardware by being memory-hard— but because litecoin uses only 128k (instead of the 8mb - 40mb discussed in the scrypt paper)  it's not really all that memory hard, I wouldn't be surprised to see LTC ASICs with _more_ efficiency gain simply because LTC-scrypt runs fairly poorly on GPUs.

But none of this is what Mike is talking about in this thread, his argument doesn't depend on the qualities of scrypt vs sha256 as a POW.  What Mike is basically saying is that if you have to make 10 chip-types rather than 1 chip type you don't get as much manufacturing scaling (e.g. you get 10x the mask costs).

Sadly, I don't think it works out this way. If you know that to be successful you'll need 10 POW implementations you'll just make a single chip that implements all ten. You'll need more of the since you're using up more area but you're back into the realm of good manufacturing scaling.   Ultimately, what counts is that the attacker isn't outspending the honest users... how you dice up the spending doesn't matter much.

Beyond that— some surprise is useful. If the attacker didn't _know_ that he was going to need 10 POWs at first, but after he sunk his costs and began his attack the honest users switched then he could indeed be caught with a useless weapon.  But now you've told him, so he knows. Good job!

There is an advantage in pooling resources so that the honest users aren't divided among multiple things— thats why merged mining was created after all.  It's a shame LTC went on the failed anti-GPU lark instead of using it.  Should such an attack happen, we could do as mike suggests— merge Bitcoin onto litecoin in order to undo LTC's mistake of dividing the resources of the honest user and making them choose... I don't see any particular advantage in doing anything in advance, however.  Public preparation would only better equip the attacker.

I am? Can you cite one of those reasons?  As far as I can tell I simply asserted it— it's obvious enough. We can't tell if the timestamps are reliable or not. That's part of the purpose of having the headline in the genesis block.  For all any of us know someone forked and replaced the first 50 blocks, or any of a zillion other things. Who knows. You've totally missed the point, but I can't figure out why. We don't know who created blocks past the first one or two. I suggested ignoring them because it gets us squarely in the realm of what we know for sure... But don't— if you like. if it suits your particular obsessions.  There weren't any conclusions being drawn from it.

And they were, compared to current software on current hardware Sergio spent a lot of time here arguing about numbers like 7.5MH/s on a single desktop. It's _laughable_.

I can't speak for DeathAndTaxes, but it's not true for me.

In fact, I don't currently own a single coin that was created prior to the date you registered on the forum (June 16, 2011) what little I'd mined before then and hadn't lost I've long since sold.

In general I'm unimpressed by your ad hominem, but I suppose I must reply to it before it gets repeated as fact in some future thread. I responded here because your posts don't make sense and because they don't fit the history as I and many others experienced it. Not because I don't want people to think Satoshi made a pretty penny— I sure hope he did: Bitcoin is a beautiful invention.

You certainly don't shy away from promoting your name or your company, or shy from making sweeping claims that might get you press... Even in this thread— I don't see a quest for knowledge. You didn't ask a question, you didn't seek people's experiences. You made a bunch of easily falsifiable assumptions (which you didn't bother to try to falsify) in order to justify a headline grabbing assertion.  I could speculate about your motivations, but I don't need to— your conclusions here are incorrect no matter what motivated you to draw them.

I'm ignoring the timestamps in the blockchain, as I said.

You might want to update the wiki.

I dunno who you're arguing with there, cause that all sounds just as nuts as Sergio's claim that he had some multiple of the hashpower required for diff 1 and but mystically the network stayed at diff 1 because??? If anything that crazy argument suggests, perhaps, that Satoshi didn't mine full time. (not that I believe it— but it's the logical conclusion of the argument that he had more hashpower than was observed in the network)

Sure, even if it was 100 machines worth that would have been true regardless— even if it took a sold 100 machines to get there.. And?

Indeed. And I've said this about a zillion times on IRC whenever someone asks about using commercial crypto accelerators for mining. ... But in all those times I'm talking about megabits not megabytes. Doh. It's easy to make doofus mistakes when they agree with the values you expect. Still doesn't get within a factor of factor of four of the numbers I was disputing, and it's still a measurement on unrelated software.  Thank you for the correction.

Ignoring??? I'm drawing what I'm saying precisely from there.  The dates in my post are from the headline in the genesis block and from the public announcement.  We don't actually know which blocks were when, because god knows how things were moved around and restructured while testing or in preparation for release.

I was actually taking Sergio's dopey argument and substituting in the data we actually had, I never said it was a good argument.  I also said _lower bound_, which you deceptively omitted in your repetition there.  I also gave a much higher number which was my measurement with contemporary hardware in software, which you've helpfully omitted.

Because I actually tested the openssl code in the reference software on a PIII that I had been using at the time and got somewhere around ~47KH/s. I'd offer to run it again but I've since moved across the country and no longer have any old computers. But heck, you won't even try the actual software— you keep _miscalculating_ (SHA256 processes 512 bytes at a time, 111 MiB/sec ~= 113KH/s of SHA256^2) based on unrelated software (and even then, it doesn't sound too inconsistent with my figures)! What the heck!  Are you allergic to actually using a computer?

By your own MIB/s numbers you're arguing that satoshi was had roughly <s>73</s>9 of your reference machines worth of computing power. (corrected per posts below)

At least the hardcopy results at 45nm were not all that impressive— a bit better than the FPGAs, but you lose all the reconfigurability. I'm doubtful that EasyPath 7 will be enough of an improvement (plus the lead time!) to really be attractive.  My thinking here is that 28nm FPGAs can have short lead times and they're reconfigurable... and this may make up for modest MH/$ and and MH/j disadvantages compares to 100nm+ mining asics.