The context is here
I don't know how exactly the new blockchain candidate was calculated in Bitcoin Network, so I guess it was based on a Merkle candidate of (a) bitcoin client/s. Also I presume that is by referring the diff of network, a client hashing the/a (golden) nonce, and then propagate it to the network, if lucky enough other nodes think this nonce is ok (the first), somewhere magic happened you get the coin (please enlighten me about this magic).
Please correct my understanding.
So (if you know Hadoop)
Despite how actually SHA(Midstate(nonce++))
was implemented in a Reduce procedure (hashing/sec/energy), how could I generate new Merkle tree(s) and midstate candidate(s) in Hadoop way (huge number of them simultaneously)?
Thank you for your reading (hard to understand).
Problem (if prior has no big problem)
What I am proposing here is, if all this works, how could a mining pool like this benefit from Hadoop? (if it could)
(If it really works, I'd implement raspberry pi to hook up hundreds of SHA asis of my own in distributed fashion.)
If it really works, i'd implement raspberry pi
Don't. It's less efficient than an Intel CPU. – Nick ODell Jul 27 '13 at 14:58