Quote from: kano on September 22, 2017, 01:55:08 PM

Quote from: mjstrand on September 22, 2017, 11:44:28 AM

Just curious if it is possible for a block to not be able to generate an answer (go through every possibility)? I wonder if anyone has analyzed how many possible answers there are for the various difficulties?

Also a tech question / idea for Kano... Are shares distributed sequentially or randomly to miners around the world? It seems that randomly picking shares may improve our luck (may make it worse too). I am a 30+ year web / c++ developer that has combined dedicated servers into a worldwide distributed digital signage solution. I have a concept on how to code random share distribution to deal with the trillions of shares. I am tempted to code up my own pool but have no idea where to start and have paying customers to keep happy. Kano, if this is distributed sequentially, can we chat?

If it wasn't possible to find a block, bitcoin would stop Tongue

Each individual hash is pseudo-random.
You'll not be able to find an association between two sequential hashes other than to say 'double-sha-256'
i.e. no.

Also, shares aren't distributed, work is.
The base work is the same for everyone on the pool, but the coinbase transaction is slightly different for everyone and somewhat sequential.
Each individual work item can produce more values to hash than the size of the difficulty, so you expect all work to be able to produce a block, but of course no one is able to actually exhaust a work item in 30 seconds unless they use a proxy to a VERY large number of miners (more than exist)

You wont be able to convince me of any effect due to the sequential client ids.
It effectively doesn't exist.

Thanks Kano. Yes, this is all super deep stuff and thanks for your time explaining it. I've only been into mining / crypto for about 2 months, so super green on it all. So based on your comment, the miner does not even complete the work that is assigned to them? That must make things insane for the server software. So does the remaining work get distributed back out to the miners again somehow or is it just ignored?

The basic explanations I have seen of mining on the web are to compute the sha256 of the block plus the number 1, then 2, then 3. Firstly, I am not certain this is even close to what truly is happening. If it is somewhat close, then I would assume that each miner is given the block info and a range of numbers to work on sequentially. I have no idea of the starting and ending range for work (and I am sure it varies by miner), but lets say the first miner's work is 1 through 1000, then my thought is rather than sending 1001 to 2000 to the second miner, the pool would send a random unprocessed starting point. For example, 1 trillion to 1 trillion + 1000. The third miner may be 1 billion to 1 billion + 1000, etc. Beyond that, each pool server would need to have their own range since the servers cannot communicate around the world fast enough. On each server, a memory cache would be filled in with the next few hundred block ranges which would get mapped into a local database in the background after being assigned and repopulate more random blocks for future requests. Of course that all gets even more insane since every miner has an individual difficulty but perhaps that could be figured in.

Eventually someone will be assigned 2001 to 3000 when it randomly comes up. Obviously if the block answer was found in the 2001 to 3000 work group, then this method dramatically slows what a sequential operation would do. Anyway, that was my thought - but it is highly possible that is already happening on every pool in the world. As I said, I'm green on this. Thanks again and it was a pleasure to chat with you and be a part of your pool for most of my short mining life.