Topic: [ANN][ANC] Anoncoin (anoncoin.net) | Privacy-centric currency | I2P darknet - page 20. (Read 40674 times)

It is normal, there is a delay of two blocks for the retargetting to understand it shall go down. If it was already going down it would have gone at the lowest allowed difficulty, but it was going up the previous block because of all the small blocks preceding.

I will try to do a delay of zero block for the retargetting, that is it will retarget the actual block through getblocktemplate if no one find it after a certain time, this is a part of the solution to resolve the problem. The second change will be in the moving averages to make it less aggressive by averagin more blocks. The last change will be the PID, I will introduce a derivative term.

All those changes were tested already on testnets but not with an hardfork, and some may be difficult to implement using an hardfork, but we will see.

Hello tomothy,

Well the network shall be alive even with only multipool hashrate around, I have some idea to make the mining more attractive but have not finish to test it, yet.

Otherwise, apparatus is ASIC Scrypt miner above 100 MH/s is fine, and solo mining or on a p2pool or on a pure anoncoin pool without autoswitcher.

Thanks

Check the first post of this thread. Mining solo with ASIC is fine. Pools are listed on the first post, but some delisted ANC recently, and all are multipools. Mining on multipools makes the problem more difficult.

There will be the release of another wallet soon, with updated PID and moving average values. Until then youcan mine using the wallet listed on the first post of this thread or master on github/anoncoin.

Thank you for this thoughtful post,

Indeed, nobody was mining for this reason. Because KGW made 150+ easy blocks when a miner found the highest difficulty block, it was a strong intensive for miners to mine after 12 h - 24 h without blocks. Then multipools automatically hoped on for the fest and difficulty went up after 150 quick blocks. Here with GR-PID the process is the same but the reward much lower as about 20 blocks are easy. So nobody care.


You are wrong, sorry in this case the diff will do like this:

Long time with no block, we are at block x:
block x -> difficulty yyy
next block x+1 -> often more difficult (delay of one block in the retargetting system, diff will go up if the diff was increasing, but down 1.7 times if it was decreasing)
block x+2 -> much easier (1.7 times easier than x+1)
block x+3 -> much easier (1.7 times easier than x+2)
block x+4 -> much easier (1.2 to 1.7 times easier than x+3) - now the average of 6 blocks down i hit and hinder the drop of diff
block x+2 -> easier (1.2 times easier than x+1) (1.2 is set by the average limit on 6 blocks and 170%, the exact value I have the data but not on hand right now)
...
~block x+20 -> difficulty at lowest because it decreased 1.2 times each time till now (if less delay in the first block then this lowest block will be ofc before x+20, and the lowest point of difficulty higher)
block x+21 -> difficulty increase x2 when the very long block is throw out of the tipfilter and is no more used in calculation of difficulty
block x+22 -> difficulty increase x2
block x+23 -> difficulty increase x1.3-x2 - now the average of 6 blocks up is hit and hinder the increase of diff
block x+24 -> difficulty increase x1.3 (1.3 is set by the average limit on 4 blocks and 200%, the exact value I have the data but not on hand right now)
...
block x+z -> difficulty reached the maximum because lots of quick blocks appeared the last 21 blocks, this was exagerated because of drop of difficulty and simultaneous increase of hashrate due to multipool switching to mine ANC. The fest is over, difficulty more difficult than LTC so multipool go away. Nobody mine the coin because there is no faithful miner.



This is complicated because there is no permanent mining going on, only multipool. Those actor dont see the advantage over KGW neither, otherwise they would mine. We need some fraction of constant hashrate.


Do not worry nobody so far is interested in updating the code but me And if someone is interested he will understand easily, the generated csv file is easy to grasp for someone motivated.


Well not dead, but certainly it has not intensive to mine like when KGW was the difficulty algo. But then it was unfair too all this instamining every 8 or 12 hours.


I dont think multipool would hop on ANC for 3 ANC. It would be interesting to try.


Not really, once the block are released on network, and 6 blocks passed, the hashrate is already a lot because the multipools are mining at this time. I would agree if nobody jumped on the chain but this is not what happen, it is just that every miner wait another miner to mine the 1 (or 2) difficult block.


Like you said user can give the fee they want, but I dont know to what extent this will motivate miner to mine the difficult block?

CS

I've tried to install the linux wallet version on a rasberry pi and am getting the error after the make command: "make -j`nproc`"

It looks like the source of the tar file has been deleted from Google and anoncoin.net?

_{user-desktop:~/anoncoin/depends$ sudo make -j`nproc`
Fetching native_protobuf...
https://protobuf.googlecode.com/files/protobuf-2.5.0.tar.bz2:
2016-09-05  ERROR 404: Not Found.
https://anoncoin.net/depends-sources/protobuf-2.5.0.tar.bz2:
2016-09-05  ERROR 404: Not Found.
funcs.mk:231: recipe for target '/home/user/anoncoin/depends/sources/download-stamps/.stamp_fetched-native_protobuf-protobuf-2.5.0.tar.bz2' failed
make: *** [/home/user/anoncoin/depends/sources/download-stamps/.stamp_fetched-native_protobuf-protobuf-2.5.0.tar.bz2] Error 8}

How can I get the make command to source the protobuf from a working source here https://github.com/google/protobuf/releases/download/v2.5.0/protobuf-2.5.0.tar.bz2 . I presume the wget url within the anoncoin qt setup needs to be modified.

Thanks for any help

please explain:
Difficulty 0.00000045
Network hash 5.65 GH/s

If you lost your pass then yes i would think so.

Is this coin making a comeback ....sad to see the original anon coin in such a sorry state really.

I would be happy to put the Prohashing pool back on Anoncoin, but the daemon just returns "error: {"code":-10,"message":"Anoncoin is downloading blocks..."}" when calling getblocktemplate even though we are on the same block as Cryptopia.

Hadn't noticed this before now after being rather lazy on the forum.

I had about 4000 anc in an old wallet for which I have lost my password. Are they now irredeemably lost because of the hard fork?

network moves again. i really don't understand that difficulty adjustment!
block 565.030 did get an even higher diff, although 565.029 took 684065 seconds to be found!

Can anyone recommend what to buy to help network? It's tough parsing through what is sensible and relatively economical (in terms of mining at a loss but not a huge loss) to buy equipment. Just wondering.

I'l willing to dedicate some hash power to this if we can get it going again. What's the latest info on pool or solo mining, software, wallet etc.?

so cryptopia is back online.. they processed deposits and withdrawals show up as transactions..

but what about the ANC blockchain itself? it is completely fucked up!
the last block was found 8 days(!) ago..

i have 24 peers, the new version 90611 and are on block 565028. cryptopia is also on block 565028.
all other block explorers are down.
so is truely nobody mining ( because of low profitability ) ?




it does even not make sense to mine now.. because of the "retarget every block", it would be the following:
block 565029 = diff 521 ( 50 % of LTC profit, so you have to pay 200% hashing price )
block 565030 = diff very low (  hashing price low but above zero )
block 565031 = diff high again, as everybody mines the low block rapidly.

so i could either mine LTC, or i could mine 1 block ANC for double price plus 1 block ANC for free. so the average value is the same. so my conclusion would be not to mine ANC.....

so i really don't see the advantage of the hardfork??


CS: you probably put a lot of work into this document ( https://wiki.anoncoin.net/GroundRod_PID_main_ver1 ).. but without a proper introduction, a defintion of the different terms and a deeper discussion of the results, its just your own brainfuck, which nobody else can understand ( nor help you with )

interpretation 1: miner perscpective
you fucked it up. ( as the chain is dead now ) - sorry

interpretation 2: user perspective
the is just no demand for a working blockchain. the peak of transactions is 43 per day!
so you could not send anything for 8 days now.. but if somebody would have a urgent transaction, he would just have to add a fee of 3 ANC , so the multipools would see high profit and process the blockchain. That would be 0.48 USD, so an affordable price for a real use case.

but again, lets look at the transaction value ( mining excluded )


interpretation 3: security
the block reward is just too low. a double-spend attack is very cheap.
 but again see point 2: nobody would care about that as nobody uses ANC:
this month there were only 25000 USD transferred. even with an optimistic fee of 0.1 % we only get 25$ for network security

Use a recent release which automatically finds peers, dont use hardcoded peers!
Also, turn on I2P.

addnodes:
213.152.161.117:55167
103.253.43.162:50438
211.149.148.149:58147
73.69.100.22:9377
88.206.141.110:55831


so why is cryptopia again "Investigating wallet issues" since a week now? ( wallet offline )

Can you please post some nodes ?

That's all CS, I just copy pasted. lol. CS continues to produce!

Thank you for the detailed summary and analysis!

Enclosed post below. Feel free to delete and make a prettier post if wanted.


GroundRod PID main ver1
Here we show the result of 5000+ blocks mining after the hardfork on block 555555 using the GroundRod PID Retarget Difficulty Algo featured in the new 0.9.6.11 version of Anoncoin.

https://github.com/Anoncoin/anoncoin/blob/master/src/pow.cpp

The PID settings used in mainet are the one from the 9.6.11 version after hardfork-block 555555.


define PID_PROPORTIONALGAIN "1.7"
define PID_INTEGRATORTIME "172800"
define PID_INTEGRATORGAIN "5"
define PID_DERIVATIVEGAIN "0"
define TIPFILTERBLOCKS_DEFAULT "21"
define USESHEADER_DEFAULT false
define NMAXDIFFINCREASE "200"
define NMAXDIFFDECREASE "170"
define DMININTEGRATOR 170
define DMAXINTEGRATOR 190
define WEIGHTEDAVGTIPBLOCKS_UP 4
define WEIGHTEDAVGTIPBLOCKS_DOWN 6
const int64_t nTargetSpacing = 180;

Here are the results so far...

It seems mining on mainet is characterized by about 10% constant hashrate and 90% multipool hashrate jumping on and off the chain. Mining quick blocks at low diff and increasing the diff until it is high enough for them to leave. The slowest block is then on hold for indefinite amount of time till the 10% constant hashrate find it, then the decrease of difficulty due to not having found a block for hours is accompanied by multipools jumping on the lower difficulty block, finding quickly many blocks while the difficulty drop and then increase to a high difficulty, which make multipools leave, repeating the cycle.

So in those very difficult mining conditions this is what we have so far at block 560555 using PID with the above settings.


Blocktime
Blocktime is too slow with around 214.6 sec instead of 180 sec, we can see when no block was found for a long time with almost vertical section and many quick blocks with almost horizontal section. The oblique parts with flatter slope are mostly due to the integral term requesting quick blocks for long times, to try to reduce the increased time toward the target of 180 sec.



Average Blocktime
This is a simple mean above all blocks since block 555149. Here we also see the avg blocktime tending to 210 sec.



Integral term
The integral term over 2 days (172800 sec) is almost always at the maximum which is equal to (190 - 180 = 10) * 5 + 180 = 230.



PIDOut
The PID output sum is characterized by periods of stability like from 1400-2000 or 3800-4400 interspersed with chaotic periods were the PIDOut achieve very high values. It is evident the higher percentage of constant mining in the stable period is enough to not have very long space without any found blocks like when the multipools alone are mining, leading to very high PIDOut due to no blocks for hours. We conclude there is some amount of constant mining relative to multipool mining which is needed to achieve stability. If nobody is mining or at a very low hashrate the high difficulty blocks then nothing can really help unless an automatically decreasing difficulty which is in fact possible with the PID (parameter useheader=1) but disabled on mainet due to lack of time consensus between peers and subsequent possible attack vector.



TipsAvg
TipsAvg is looking quite OK but it hides information about very slow blocks followed by very quick blocks.


Space
Block spacing shows the very long blocks that anoncoin is subject to. This is no new to PID, very long blocks were also found with KGW due to the same multipools inconstant hashrate. The difference between KGW and PID is PID gives much less quick block than KGW which gave around 140 at a time.



Space zoom


A zoom on the space chart shows the very quick blocks following the very long blocks. This is of course due to conjugated effect of decrease of difficulty by the PID to remove the impact of the long block and increased hashrate due to multipools jumping onto the chain to get the easy difficulty blocks.

Log of difficulty
A logarithm of difficulty shows the extent of variation of difficulty between the slow difficult block and the quick easy blocks. We can see there is two order of magnitudes between those two extremes which is a lot. This is partly due to a too strong retargetting of the difficulty with these particular proportional gain, maximum diff increase and decrease and the too short periods of 4 and 6 blocks for both moving averages.




Difficulty chart
The difficulty chart shows 400 blocks using the old Kimoto Gravity Well retarget algo followed by about 5000 blocks using the new GroundRod PID.

It is very well seen on this chart that sometimes there is only multipools mining with no constant hashrate (p.ex. 749-1378) and other times there is a constant hashrate stabilizing the fluctuation (p.ex. 1378-2092). The same conclusion can be done which is Anoncoin need some fraction of constant mining to have a well behaving chain, but miners have more incentive to mine on multipools so it is not easy to get this constant hashrate.




Conclusion

The GroundRod Retarget PID was tuned for quick retargetting in case of high fluctuation of difficulty, but with the necessity to have at least ca one hundred blocks at each hashrate levels. Here the pattern of mining is more like mine 25 easy, let others mine one hard, mine 25 easy, let others mine one hard. Consequently I am in the process of tuning the parameters of the PID to have less aggressive variation of difficulty for then the ratio between easy and hard difficulty block will be less than one hundred as it is actually, limiting the impact of higher and lower difficulty and hopefully smoothing the block distribution. The community will be updated when I will be satisfied with the new parameters. CS

AFAIK Meeh has not been active with ANC for a long time. I assume that his failure to even update the OP of the announcement thread regularly was part of the reason you created this new thread.

Based on his long absence why are you still posting content on a site hosted on a server he controls? It seems to me that copying the important content and transferring it to a new site under your control would be a better option. For a long time you have been far more reliable than Meeh in providing updates to the community and answering questions.

You have done a lot for ANC. I think some of the old community might return if things continue to move in the right direction.