|
Title: sfh Post by: Anonymous on November 16, 2010, 01:03:57 AM adg
Title: Re: Are you able to tell when the chain temporarily forks? Post by: theymos on November 16, 2010, 01:08:37 AM Is there any way I can confirm, that someone else generated a block for the same block as me, and then it forked, and the other chain happened to grow longer faster? Search debug.log for "REORGANIZE" right after your "proof-of-work found". That means another chain has gained the lead over your chain. That's why we have this... Code: // Don't show generated coin until confirmed by at least one block after it Title: Re: Are you able to tell when the chain temporarily forks? Post by: theymos on November 16, 2010, 01:30:36 AM Yikes, I've got it 4 times in there in the last couple of days. Is that each time I found one and it got longer, or anytime I hopped on the shorter chain? You're more likely to reorganize right after starting Bitcoin while you catch up with the current blocks. If you were running Bitcoin that whole time, I'm not sure why you would get so many. It only happens when you stop working on one branch and start working on another one. I would be interested in seeing that 24-hour section of your debug.log. Title: Re: Are you able to tell when the chain temporarily forks? Post by: Anonymous on November 16, 2010, 10:51:01 AM Yikes, I've got it 4 times in there in the last couple of days. Is that each time I found one and it got longer, or anytime I hopped on the shorter chain? You're more likely to reorganize right after starting Bitcoin while you catch up with the current blocks. If you were running Bitcoin that whole time, I'm not sure why you would get so many. It only happens when you stop working on one branch and start working on another one. I would be interested in seeing that 24-hour section of your debug.log. It was actually as of Friday, so a couple of days. I'll send it if you'd like. Further investigation shows it found 2 blocks, and after each of those it had a reorg in there. Plus two others. One, right near the start, and one earlier today. Definitely weird. Burns me that it looks like I found two but they weren't accepted, darn it haha. I hate that feeling when you scratched a scratch lotto ticket and think you won only to realise the last number needed an extra zero. :( Title: Re: Are you able to tell when the chain temporarily forks? Post by: ByteCoin on November 18, 2010, 12:57:28 AM Are there really so many hashes found that go to waste?
The quote in the comment in ui.cpp says that by design something in the region of 10% of blocks generated won't be accepted, presumably because another block has been found slightly before and is propagating across the network but hasn't reached our computer yet. For about 10% of blocks to be wasted the propagation time for a block from one generator to another would have to be of the order of about a minute! Surely block propagation can't take that long? Could some keen miner (Artforz) please scour their logfile and find out what the ratio of accepted to rejected blocks is in real life? Also, I notice that it's in the miner's mutual best interests (jointly and severally) to ensure that they connect directly with each other and not with or via non-generating clients. It's most important to connect to generators with the highest hash rate. In fact, there's no incentive for generators to talk to non-generators at all unless non-generators are offering a transaction with a fee! ByteCoin Title: Re: Are you able to tell when the chain temporarily forks? Post by: FreeMoney on November 18, 2010, 01:20:52 AM Are there really so many hashes found that go to waste? The quote in the comment in ui.cpp says that by design something in the region of 10% of blocks generated won't be accepted, presumably because another block has been found slightly before and is propagating across the network but hasn't reached our computer yet. For about 10% of blocks to be wasted the propagation time for a block from one generator to another would have to be of the order of about a minute! Surely block propagation can't take that long? Could some keen miner (Artforz) please scour their logfile and find out what the ratio of accepted to rejected blocks is in real life? Also, I notice that it's in the miner's mutual best interests (jointly and severally) to ensure that they connect directly with each other and not with or via non-generating clients. It's most important to connect to generators with the highest hash rate. In fact, there's no incentive for generators to talk to non-generators at all unless non-generators are offering a transaction with a fee! ByteCoin Is the fact that they might send a fee or might generate enough? What is the real cost of connecting? Why not just keep connecting to more and more people? |