Mtgox Yubikey and storage

[Kluge]

This will never happen.

Excellent... ummm... how does that work then?

It's a probability thing.  THe probability of it happening is so low as to be negligible.

Even if you could generate a trillion addresses every second for the rest of your life, generating someone else's address is still less probable than the chance that Barack Obama will pick you up as a hitchhiker, take you to a gas station, where you'll buy the winning lottery ticket for Mega Millions at 12:12pm on 12/12/12 with a ten-dollar bill that fell out of an overhead helicopter into your shirt pocket while the Pope is simultaneously being struck by lightning.

How long would it take the super-ist super-computer to generate & save 1/10 of all possible addresses? .0686 btc bounty!

[gohan]

How long would it take the super-ist super-computer to generate & save 1/10 of all possible addresses? .0686 btc bounty!

I'll give it a shot, someone please correct me if I'm wrong.

Let's say we generate 10 trillion addresses per second, which corresponds to a few times more than 100,000 high-end ATI cards.

2^160 / 10 / 10^13 / 60 / 60 / 24 / 365 / 1000, or ~ 4.63x10²³ is the number of millennia it would take to generate nearly 1/10 of all possible addresses. "Nearly", because we disregarded collisions.

There are almost 2²⁵⁶ (2²⁵⁶-2³²-2⁹-2⁸-2⁷-2⁶-2⁴-1) valid private keys. Two private keys can correspond to the same address. So actually it would take ~ 3.67x10⁵² millennia to be absolutely sure you've generated 1/10 of all possible addresses.

If it's possible to check whether you have generated an address before, with zero cost, then you will know that you have generated 1/10 of all possible addresses some time between 4.63x10²³ and 3.67x10⁵² millenia. By comparison, the universe existed for about 1.4x10⁷ millennia.

[casascius]

How long would it take the super-ist super-computer to generate & save 1/10 of all possible addresses? .0686 btc bounty!

My short answer would be that if hard drives could each hold a million terabytes, it would still take a batch of hard drives greater than the weight of the entire earth to save all of those addresses.

[memvola]

How long would it take the super-ist super-computer to generate & save 1/10 of all possible addresses? .0686 btc bounty!

My short answer would be that if hard drives could each hold a million terabytes, it would still take a batch of hard drives greater than the weight of the entire earth to save all of those addresses.

According to my back of the envelope calculation, that would be 304479.5 earth masses (~ 1 solar mass). (Assuming 1 exabyte per 500 gr. hard drive, each address is 25 bytes.)

[Kluge]

How long would it take the super-ist super-computer to generate & save 1/10 of all possible addresses? .0686 btc bounty!

I'll give it a shot, someone please correct me if I'm wrong.

Let's say we generate 10 trillion addresses per second, which corresponds to a few times more than 100,000 high-end ATI cards.

2^160 / 10 / 10^13 / 60 / 60 / 24 / 365 / 1000, or ~ 4.63x10²³ is the number of millennia it would take to generate nearly 1/10 of all possible addresses. "Nearly", because we disregarded collisions.

There are almost 2²⁵⁶ (2²⁵⁶-2³²-2⁹-2⁸-2⁷-2⁶-2⁴-1) valid private keys. Two private keys can correspond to the same address. So actually it would take ~ 3.67x10⁵² millennia to be absolutely sure you've generated 1/10 of all possible addresses.

If it's possible to check whether you have generated an address before, with zero cost, then you will know that you have generated 1/10 of all possible addresses some time between 4.63x10²³ and 3.67x10⁵² millenia. By comparison, the universe existed for about 1.4x10⁷ millennia.

Additional .6174BTC bounty for percent chance of that super-ist super-computer creating an address which collides with any already-existing address within 20 years.
+.314BTC bounty to give percent chance of successfully brute-forcing this password ("ingk32%(#ldNUI93") on a WinZip archive with AES-128 encryption within 1 minute using 100k of those same ATI cards, assuming they are each able to push out 4million valid passwords per second.

ETA: Feel free to round to the nearest billionth.

Additional .6174BTC bounty for percent chance of that super-ist super-computer creating an address which collides with any already-existing address within 20 years.

https://bitcointalk.org/index.php?topic=4858.msg71102#msg71102
It is getting off-topic though, why don't you create a new thread to resume this? Large numbers kinda turn me on I guess...

Heh. Forgot what this thread was even about. I'll quit bumping.

[memvola]

Additional .6174BTC bounty for percent chance of that super-ist super-computer creating an address which collides with any already-existing address within 20 years.

https://bitcointalk.org/index.php?topic=4858.msg71102#msg71102

It is getting off-topic though, why don't you create a new thread to resume this? Large numbers kinda turn me on I guess...