[cryptography] bitcoin stats (Re: OT: Skype-Based Malware Forces Computers into Bitcoin Mining)

Adam Back adam at cypherspace.org
Thu Apr 18 04:19:43 EDT 2013

vs the other things malware does it also seems like a much more benign
"payload" - uses a bit more electricity!  I imagine they throttle it down
dynamically when the user actually does things to hide the computer slow
down.  (Though typically you wont notice with GPU mining unless you are
playing video games).  But you may notice the increased GPU fan noise and
heat on mid/high range cards.

Seems bitcoin mining is using 40MW of power (estimate) thats 0.025% of the
amount used by US households (30,000 US households - obviously bitcoin
mining is global - its just a comparative stat).

Thats a lot of 55.1 bit hashcash mining those miners are doing!  Bits is
more human readable by my thinking from hashcash difficulty measurement and
visually easy to confirm from the hash output.


you can see the SHA256 output 


has 13 hex 0 nibbles = 52 bits plus the next digit is a one 0001 in binary
so thats three more - so its a 55 bit.

And the difficulty parameter is measured in the number of 2^32 hashes
(MAX_UINT == ~4 billion) so you can get back to it from 55.1-32=23.1 via
2^23.1 (in bc e(l(2)*23.1) ~ 9 million.  Much better :)  And converting GPU
card mining stats into bits is also easy 400Mh (amd 7870) = 28.6.  So that
means I have 55.1-28.6 = 26.5 bits short (I need 2^26.5 bits to match the
network hash rate).  And also you can recreate the network hashrate as
2^55.1/600 (60 seconds x 10 mins per block) or subtract 9.2 as log2(600) =
9.22 so 55.1-9.2=45.9 = 2^45.9 so subtract 40 for terahash/sec = 2^5.9 =
59.7.  (Except some web stats are misreporting network hash in 1000^4 rather
than 1024^4).

Anyway I think humans work better with massive numbers in the log scale.


On Wed, Apr 17, 2013 at 10:27:01PM -0400, Kevin W. Wall wrote:
>You know Bitcoin must have "arrived" when this is going on.
>(For that matter, I even heard Bitcoin mentioned on NPR a few
>days ago.)
>As reported on IEEE Computer Society's _Computing Now_
>news site:

More information about the cryptography mailing list