Open educatedwarrior opened 10 years ago
I have this feature in the mintypool.com code, along with a bunch of other metrics to support splitting hash to more than one coin.
AFAIK there isn't a way to get the blocker time from the wallet via RPC, if there is please let me know!
I think it would have to be a calculation of difficulty vs pool hash rate would give you a close estimation of block time.
On Sunday, June 8, 2014, Jerry Brady notifications@github.com wrote:
I have this feature in the mintypool.com code, along with a bunch of other metrics to support splitting hash to more than one coin.
AFAIK there isn't a way to get the blocker time from the wallet via RPC, if there is please let me know!
— Reply to this email directly or view it on GitHub https://github.com/zone117x/node-open-mining-portal/issues/278#issuecomment-45442041 .
Right, but the idea is that if you also know the target block time of the coin itself you can optimize to prevent generating orphaned blocks. If your hashrate hits a block every 2 minutes and the coin's target time is 10 min for instance.
This sounds kinda like Gambler's fallacy. I don't think it mathematically matters. If your pool is has statistically mined 90% (for example) of the time it takes to find a block, you aren't going to "lose" that effort spent mining if you switch to another coin.
maybe i don't understand the block generation process well enough then - because i thought everyone on the network was competing to solve the current block and once that happened, everyone's work started over.
If some coin has a difficulty to where statistically it would take 10 minutes of 100MH/s hashing (for example) to solve the block, you can spend that 10 minutes hashing and still not find a block. You can spend 10 hours mining and if you were very unlucky and didn't solve a block, you still aren't any more likely to solve the block then you were from when you first started 10 hours ago.
Block target times are just a probability, not a linear progress sort of thing.
When a block is solved on the network then pools/workers don't really "start over" they just start hashing at different data. It also doesn't take a block being solved to change the data being hashed, for example, new transactions can trigger new work being sent to the miners.
Go read up about Bitcoin "Luck"
You either hit a block or you dont, the powers not wasted if it switches the shares for that round are stored, when the pool returns to mine those shares are still used to calulate the payout.
You dont loose anything buy switching midblock
The switch block
If you set up 10 minute switching for Coins with long block targets (like Curecoin which has a 10 minutes block target) you run the risk of switching to a coin and having a very low probability of finding a block in the set switching period.
If there was an option to add an additional condition for switch so that the coin switch would occur if the time range has occurred and a block was found, that would mitigate against using pool switching on coins with long block targets.