ColdHardMetal
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 23, 2010, 05:30:50 PM |
|
Tried cooperative mining last night, had a few issues: - Started my first worker, after an hour or so I had managed to send back two shares, things were looking good.
- Reading on the forum I found that if I am going to start another worker, I should register a new worker, so I did that, and then the shares I used to have went back to zero!
- Started up the workers anyways, and left them running all night
- Accumulated a couple more shares over night, but then this morning the worker segfaulted, and suddenly my shares are back to zero according to my profile
I like the idea of cooperative mining, but something doesn't seem right with how the statistics are gathered if I can easily loose things like this. When the pool finds a block, coins are allocated based on your number of shares, and then share values get set back to zero for the next block. Do you have any values in the "Unconfirmed Reward" or "Reward" boxes?
|
|
|
|
hacim
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 64
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 23, 2010, 07:05:04 PM |
|
Tried cooperative mining last night, had a few issues: - Started my first worker, after an hour or so I had managed to send back two shares, things were looking good.
- Reading on the forum I found that if I am going to start another worker, I should register a new worker, so I did that, and then the shares I used to have went back to zero!
- Started up the workers anyways, and left them running all night
- Accumulated a couple more shares over night, but then this morning the worker segfaulted, and suddenly my shares are back to zero according to my profile
I like the idea of cooperative mining, but something doesn't seem right with how the statistics are gathered if I can easily loose things like this. When the pool finds a block, coins are allocated based on your number of shares, and then share values get set back to zero for the next block. Do you have any values in the "Unconfirmed Reward" or "Reward" boxes? Ah I do have something in "Unconfirmed Reward", so that is probably it... Although its still strange that the counts went to zero as soon as I registered a new worker, it could be a coincidence, but that seems unlikely. Also, the segfault sucked.
|
15yns1RVpBHZ8uj8mGVUJVCyPh5ieW3FQx
|
|
|
DiabloD3
Legendary
Offline
Activity: 1162
Merit: 1000
DiabloMiner author
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 24, 2010, 09:31:03 PM |
|
I suggest everyone who uses my miner to upgrade to the newest version, it should fix the final issue with the pool rejecting shares.
|
|
|
|
slush (OP)
Legendary
Offline
Activity: 1386
Merit: 1097
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 25, 2010, 01:19:18 PM |
|
- Reading on the forum I found that if I am going to start another worker, I should register a new worker, so I did that, and then the shares I used to have went back to zero!
Everytime pool find a block, all shares are set to 0 for next round. Adding new workers cannot affect it. - Accumulated a couple more shares over night, but then this morning the worker segfaulted, and suddenly my shares are back to zero according to my profile
Shares dropped to zero because next block was found ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) , miner segfault cannot affect it, too. But I believe it is very annoying. Please report all crashes to miner developer (you are using CPU miner, so report to jgarzik, please). Bug reports can help a lot with miner stability, because developers cannot test their software on all types of platforms. I like the idea of cooperative mining, but something doesn't seem right with how the statistics are gathered if I can easily loose things like this.
No one share was lost during your 'problems'. Please read instruction on mining.bitcoin.cz homepage, there is described handling with shares.
|
|
|
|
LZ
Legendary
Offline
Activity: 1722
Merit: 1072
P2P Cryptocurrency
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 25, 2010, 01:20:52 PM Last edit: January 22, 2018, 09:25:49 AM by LZ |
|
|
My OpenPGP fingerprint: 5099EB8C0F2E68C63B4ECBB9A9D0993E04143362
|
|
|
slush (OP)
Legendary
Offline
Activity: 1386
Merit: 1097
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 25, 2010, 02:07:19 PM |
|
I thought puddinpop's miners are using 'proprietary' protocol (not compatible with getwork() from bitcoin RPC). I will test it on my dev environment. If it really works with pool, I'll add them to homepage.
|
|
|
|
slush (OP)
Legendary
Offline
Activity: 1386
Merit: 1097
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 25, 2010, 09:05:49 PM Last edit: December 25, 2010, 09:19:00 PM by slush |
|
After many days looking at pool stats, I see that many users which succesfully submitted winning hashes suddenly disconnected from pool. I asked few of them and their response was (in short) 'I was lucky, I found three blocks, but my pool reward was deeply under 100BTC. I don't need the pool'. Pooled mining works exactly like this; it is fair in middle/long term, not in short term. Sometimes is lucky John (and he sponsors others) and sometimes is lucky Pepa. Nobody knows when he will be lucky and when he will need 'support' from others. Personally I understand this demotivation, because strong players who contribute 100s mhashes are de facto sponsoring tens of others for long time, until others find a valid block (but it already happen, pool has blocks also from relative poor workers). This leads me to the idea which Satoshi described in his earlier post in this thread: the user who submits the hit that solves the block should get an extra amount off the top, like 10 BTC.
Originally, I didn't understand, why somebody should get 'premium' just because he was lucky and submit valid block. Now I'm changing my mind; it is not about probability and statistics, but also about psychology and motivation. With current rules, the biggest motivation is between slow miners, because it is their only way how to make a reward in finite time; but without strong players, there won't be thousands of mhashes in pool and no daily rewards for all participants. I'm thinking about adding 10% (5BTC) from block as 'premium' to block founder. This should solve not only motivation of strong players to contribute to pool, but also possible problem with pool sabotages (not submitting valid block by false miner); when saboteur miss this premium, it is much bigger loss than only miss reward from share calculations. So adding 'premium reward' into pool economy should make pool stronger in two ways at once. Final formula for calculating reward would be: (user's shares / total shares) * 45BTC for all participating in round; next 5 BTC for block founder.This should motivate players which pool needs and should not make big loss for others (it is only 10% less in reward if user never find valid block, which is unlikely for loyal miners) I'm only thinking about it, I'm definitely not sure with this and I would like to discuss it with you, pool users. Please comment it and tell me your opinion. I'm also sorry for mostly incoherent content in my pseudoEnglish ![Grin](https://bitcointalk.org/Smileys/default/grin.gif)
|
|
|
|
slush (OP)
Legendary
Offline
Activity: 1386
Merit: 1097
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 25, 2010, 09:25:23 PM |
|
I suggest everyone who uses my miner to upgrade to the newest version, it should fix the final issue with the pool rejecting shares.
Thank you Diablo for many hours of debugging and fixing this strange bug. I confirm that latest version is working very well and no shares are missing anymore. All of you using Diablo's miner, please update. It should slightly improve your shares in pool. Also thanks for all proposals for new service name :-D. I renamed it to "Bitcoin Pooled Mining", BPM is AFAIK not related to any controversal topic :-). I also want to have 'pool' and 'mining' inside to not confuse newcomers.
|
|
|
|
dsg
Jr. Member
Offline
Activity: 37
Merit: 2
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 12:04:26 AM Last edit: December 26, 2010, 12:49:52 AM by dsg |
|
I'm thinking about adding 10% (5BTC) from block as 'premium' to block founder.
I'm against this personally. It seems to upset the fairness of the pool (and goes against the point, to make it practical for slow miners to participate). But I'm willing to listen to arguments otherwise. In my opinion it's also not too big a problem if there isn't a payout every day. The pool ensures return from mining of hardware that would take months or years to generate. What's the problem with having to wait 2-3 days for the pool to find a block? So my vote is no to the changes, the payout should be purely dependent on contribution (shares). If it matters I am contributing 175-195Mh/s to the pool.
|
|
|
|
nikitakit
Newbie
Offline
Activity: 3
Merit: 0
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 08:21:09 AM |
|
There seems to be a lot of mentions of "shares" here. Can someone tell me what a share is and what you need to get one? I'm running some miners on my computer, but my share count remains at 0. I'm running self-compiled cpuminer-0.3.2 in the 1000-2000 khash/sec range.
|
|
|
|
ColdHardMetal
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 09:16:00 AM |
|
There seems to be a lot of mentions of "shares" here. Can someone tell me what a share is and what you need to get one? I'm running some miners on my computer, but my share count remains at 0. I'm running self-compiled cpuminer-0.3.2 in the 1000-2000 khash/sec range.
Some of us have pooled our processing power in a shared effort to find blocks. For each chunk of a block (that may be a gross oversimplification/misunderstanding on my part of what exactly is going on in terms of block generation) your machine(s) completes you get a share of the 50 coins awarded when the pool completes a block. For example if it takes processing of 5000 chunks to find the next block and your machine processed 20 of them, then you get (20/5000)*50 = .2 coins. Check out the OP in this thread if you want to join the effort.
|
|
|
|
slush (OP)
Legendary
Offline
Activity: 1386
Merit: 1097
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 09:25:04 AM |
|
There seems to be a lot of mentions of "shares" here. Can someone tell me what a share is and what you need to get one?
In terms of pooled mining, share is one 'solved block' by your miner on very low difficulty. Your reward from pooled mining is adequate to "your shares / total shares * 50". More shares, more reward. I'm running some miners on my computer, but my share count remains at 0. I'm running self-compiled cpuminer-0.3.2 in the 1000-2000 khash/sec range.
Looks like you are not connected to pool. Did you provide URL and user/pass on miner commandline? On 1000-2000khash machine you should see one solved block (=share) in ~1 hour.
|
|
|
|
nikitakit
Newbie
Offline
Activity: 3
Merit: 0
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 09:43:38 AM |
|
I have two miners, with slightly different configurations. (The difference of space vs. equals sign is a problem on my part, but I don't see it affecting the argument interpretation). ./minerd --url http://mining.bitcoin.cz:8332 --userpass nikitakit.cpu1:*** --threads 2 --algo cryptopp_asm32 ./minerd --url=http://mining.bitcoin.cz:8332 --userpass=nikitakit.cpu2:*** --algo=cryptopp_asm32 I just rebooted the second one with protocol debugging (-P), and it showed some JSON responses.
|
|
|
|
slush (OP)
Legendary
Offline
Activity: 1386
Merit: 1097
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 09:50:24 AM |
|
I just rebooted the second one with protocol debugging (-P), and it showed some JSON responses.
I see your workers in system, but they have no shares yet. Please wait until miners show you found block. Then shares should be something >0. I wrote that you should have one block per hour, but of course it is still probabilistic. Maybe you are unlucky and have to wait much more for your first share.
|
|
|
|
nikitakit
Newbie
Offline
Activity: 3
Merit: 0
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 10:03:37 AM |
|
I guess it's that my system is slow (the 1-2K is the sum of the threads, not the individual ones). I get the "zeroes in hash" message once in a while, but it keeps telling me that it's not the actual result.
I'm still completely confused about the block/share system. The way I understand it is this: The goal of the system is apparently to find a block, after which point coins are awarded and shares reset. But then how can 1 share = 1 block? I'm really new to bitcoin, so I don't understand most of the terminology used here.
|
|
|
|
doublec
Legendary
Offline
Activity: 1078
Merit: 1005
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 10:37:39 AM |
|
I guess it's that my system is slow (the 1-2K is the sum of the threads, not the individual ones). I get the "zeroes in hash" message once in a while, but it keeps telling me that it's not the actual result.
I'm pretty sure that cryptopp_asm32 is broken on at least some systems (mine included). It never generates a share. There is some mention of this in the thread about minerd. I stick to the 'C' method or '4way' on compatible hardware.
|
|
|
|
slush (OP)
Legendary
Offline
Activity: 1386
Merit: 1097
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 07:12:18 PM |
|
I'm thinking about adding 10% (5BTC) from block as 'premium' to block founder.
Well, during last day I talked with few people and the most of you don't like the idea, even you with strong GPUs in pool. So I decided to let it be as it is; every share has the same price and pool does not divide workers to 'strong' and 'poor' in any way.
|
|
|
|
slush (OP)
Legendary
Offline
Activity: 1386
Merit: 1097
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 07:18:26 PM |
|
I guess it's that my system is slow (the 1-2K is the sum of the threads, not the individual ones). I get the "zeroes in hash" message once in a while, but it keeps telling me that it's not the actual result.
doublec has good point; please try miner with another algo and report your crash to jgarzik. I'm still completely confused about the block/share system. The way I understand it is this: The goal of the system is apparently to find a block, after which point coins are awarded and shares reset. But then how can 1 share = 1 block? I'm really new to bitcoin, so I don't understand most of the terminology used here.
In term of pool, one share is 'block' with very low difficulty (terms 'block' and 'difficulty' are not related to pool). When you connect miner to pool, your miner thinks he is solving real blocks, but he is solving those low difficulty blocks (shares). So when miner tell you 'block found', it does not mean full block (50 BTC), but only one share in pool. But time to time this 'share' is also valid block for Bitcoin network (reasons behind it are bit technical), so when worker submit share which is also block, 50 BTCs from Bitcoin network are divided between users using formula, which is described on pool homepage. I hope it is cleaner to you now.
|
|
|
|
Rai
Newbie
Offline
Activity: 12
Merit: 0
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 08:31:04 PM |
|
I guess it's that my system is slow (the 1-2K is the sum of the threads, not the individual ones). I get the "zeroes in hash" message once in a while, but it keeps telling me that it's not the actual result.
I'm pretty sure that cryptopp_asm32 is broken on at least some systems (mine included). It never generates a share. There is some mention of this in the thread about minerd. I stick to the 'C' method or '4way' on compatible hardware. I can confirm this. I was running cryptopp_asm32 on my laptop here and getting false proofs all the time. I switched it to 4way and my first proof came back true. It's running much slower than cryptopp_asm32, but it's working at least.
|
|
|
|
nanotube
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
December 26, 2010, 10:47:13 PM |
|
<snip> I hope it is cleaner to you now.
I have written up a description of your system on the wiki a few days ago: https://en.bitcoin.it/wiki/Pooled_Mininghope it saves you some explaining time in the future. ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif)
|
|
|
|
|