jch9678
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 03:25:09 PM |
|
Again, statistically it doesn't matter, it's just an issue with people that want to always get paid even for small periods of time. If a coin is profitable, it still is regardless of difficulty.
you have your math - i have my practical stats, and it says - don't mine when those coins are on yaamp, both too fast or too slow, if you want to mine those coins themselves - go to coin's pool or solo, not yaamp that is all good and well if that is YOUR opinion ... especially when considering yaamp.com is actually gone ... it is no more ... the pool that is being discussed at the moment is a NEW pool ... how does that make mining at another pool or solo any different to mining using that pool that is being discussed ... its not yaamp.com - its just using the code ... #crysx i mean yaamp's code, not site, just try and see for yourself, multialgo switching makes huge difference i see what you mean ... on this fact alone - it does seem that way ... multialgo switching has abrupt changes to contend with and so suffer share rejects everytime the switch happens ... this is a loss of shares and productivity - even though they are small between switches ... we have seen this happen with our own miners / farm ... but that is with ANY multialgo multipool - not just yaamp.com ... look at epsylon3 .. he is one hell of a developer - and now running his own yaamp-coded pool with more algos added ... it sort of goes to show that if even a seasoned developer like him is willing to look at - setup - and improve the code - then the pool code is still on a reasonably good level ... nicehash has the same sort of issues ( though they dont mine coins - just divert hashrate ) with multialgo switching ... so much so - that they devote an entire section on the website to it ... single algo / single coin pools may be more 'efficient' ( for lack of a better word ) at mining - but we believe that the losses are so minimal thats its close to negligible ... #crysx We are not aware of latest sgminer having any issues with algo switching. There were issues at start, but these times are long gone. If you use any other software besides sgminer and getting rejects, then there is an issue with that software. you are quite right ... it is the earlier version ( 5.1.0-dev sgminer ) while using the amd cards - but mostly it is ccminer ( spmod fork in this case - as nvidia cards are the main miners we have in our farm ) ... using just sgminer on ONE algo ( quark in this case ) also has a large number of share rejections ... testing the latest that we have compiled ( sgminer 5.1.1.17-g6666 ) still manages to reject shares on the stratums ... the stratums we have had the pleasure of testing here are x11 - x13 - x15 - neoscrypt - lyra2re - quark ... currently - the share reject rate is nominal ( ie - small ) but at times jump into astronomical figures for short periods of time ... we have a number of miners currently running on quark using your us-based startum ( westhash ) with both ccminer ( spmod fork 1.5.53-git - compiled under linux x64 and cuda 6.5 - ) and the russian sgminer miner ( untrusted but works - 5.1.1 - under windows ) as well as the latest sgminer git compile in linux ( sgminer 5.1.1-17-g666 ) and ALL of them show share rejects on a regular basis mining on quark.usa.nicehash.com:3345 ... note - this is mining 'just' quark ... not algo switching ... i have seen sgminer ( git ) show regular share rejects switching algos also - though the testing is no where near as comprehensive as what we have already done on each algo singularly ... take a look at the current stats on the test bench as we chat - https://www.westhash.com/?p=miners&a=12&addr=15umzHXF8NzXA4FywmeFbrDHgL8WcPs3wx ... we are working alongside another long term respected user here on bct ( who we wont name for the moment ) on an algo switching script for linux using nicehash and ( originally ) yaamp ... obviously nicehash is the main testbed now until a new yaamp comes about ( which is currently planned ) - but these tests show proof that the farm in all its different incarnations and hardware / software makeup will never submit shares that are ALWAYS accepted ... this is not to single out any one pool or another ... it is to show ( at least to us ) that rejected shares can only be minimized - not altogether eradicated ... no matter how 'good' the mining software - or mining hardware - or stratum software / pool software is ... so even though you may not be aware of the share rejects - our mining proves otherwise ( at least for the miners we have tested internally ) ... we are putting together all the donation links through nicehash - which means that even though share rejects happen - it is almost negligible in the long run ... btw - havent heard back from support whether the other ip addresses we have submitted ( for the donation links ) have been whitelisted yet ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) ... #crysx Strange, I'm using the Russian Miner at westhash and I have zero rejects over a long enough period of time to be significant. . Literally zero, I've never seen anything like it. I'm much closer to the server than you are, maybe it's the latency that's causing the rejects.
|
BTC: 15GqpmqNNJ1REWrDWTfymh7moos1sEvz7A
|
|
|
joblo
Legendary
Offline
Activity: 1470
Merit: 1114
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 03:34:05 PM |
|
i mean yaamp's code, not site, just try and see for yourself, multialgo switching makes huge difference
i see what you mean ... on this fact alone - it does seem that way ... multialgo switching has abrupt changes to contend with and so suffer share rejects everytime the switch happens ... this is a loss of shares and productivity - even though they are small between switches ... we have seen this happen with our own miners / farm ... but that is with ANY multialgo multipool - not just yaamp.com ... look at epsylon3 .. he is one hell of a developer - and now running his own yaamp-coded pool with more algos added ... it sort of goes to show that if even a seasoned developer like him is willing to look at - setup - and improve the code - then the pool code is still on a reasonably good level ... nicehash has the same sort of issues ( though they dont mine coins - just divert hashrate ) with multialgo switching ... so much so - that they devote an entire section on the website to it ... single algo / single coin pools may be more 'efficient' ( for lack of a better word ) at mining - but we believe that the losses are so minimal thats its close to negligible ... #crysx I don't have a problem with rejects upon switching algos on any multipool. I don't use the pools' multialgo features, but I also incur the overhead of killing and restarting ccminer. I don't think there is a truly elegant way to switch algos. The pool signals by disconnecting so any unsubmitted shares are lost. Whether the miner reports that as a reject is irrelevant, it's lost work. To have a seamless algo switch requires a cooperative disconnect where the miner is allowed to complete the work in progress.
|
|
|
|
chrysophylax
Legendary
Offline
Activity: 2870
Merit: 1091
--- ChainWorks Industries ---
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 03:46:38 PM |
|
Again, statistically it doesn't matter, it's just an issue with people that want to always get paid even for small periods of time. If a coin is profitable, it still is regardless of difficulty.
you have your math - i have my practical stats, and it says - don't mine when those coins are on yaamp, both too fast or too slow, if you want to mine those coins themselves - go to coin's pool or solo, not yaamp that is all good and well if that is YOUR opinion ... especially when considering yaamp.com is actually gone ... it is no more ... the pool that is being discussed at the moment is a NEW pool ... how does that make mining at another pool or solo any different to mining using that pool that is being discussed ... its not yaamp.com - its just using the code ... #crysx i mean yaamp's code, not site, just try and see for yourself, multialgo switching makes huge difference i see what you mean ... on this fact alone - it does seem that way ... multialgo switching has abrupt changes to contend with and so suffer share rejects everytime the switch happens ... this is a loss of shares and productivity - even though they are small between switches ... we have seen this happen with our own miners / farm ... but that is with ANY multialgo multipool - not just yaamp.com ... look at epsylon3 .. he is one hell of a developer - and now running his own yaamp-coded pool with more algos added ... it sort of goes to show that if even a seasoned developer like him is willing to look at - setup - and improve the code - then the pool code is still on a reasonably good level ... nicehash has the same sort of issues ( though they dont mine coins - just divert hashrate ) with multialgo switching ... so much so - that they devote an entire section on the website to it ... single algo / single coin pools may be more 'efficient' ( for lack of a better word ) at mining - but we believe that the losses are so minimal thats its close to negligible ... #crysx We are not aware of latest sgminer having any issues with algo switching. There were issues at start, but these times are long gone. If you use any other software besides sgminer and getting rejects, then there is an issue with that software. you are quite right ... it is the earlier version ( 5.1.0-dev sgminer ) while using the amd cards - but mostly it is ccminer ( spmod fork in this case - as nvidia cards are the main miners we have in our farm ) ... using just sgminer on ONE algo ( quark in this case ) also has a large number of share rejections ... testing the latest that we have compiled ( sgminer 5.1.1.17-g6666 ) still manages to reject shares on the stratums ... the stratums we have had the pleasure of testing here are x11 - x13 - x15 - neoscrypt - lyra2re - quark ... currently - the share reject rate is nominal ( ie - small ) but at times jump into astronomical figures for short periods of time ... we have a number of miners currently running on quark using your us-based startum ( westhash ) with both ccminer ( spmod fork 1.5.53-git - compiled under linux x64 and cuda 6.5 - ) and the russian sgminer miner ( untrusted but works - 5.1.1 - under windows ) as well as the latest sgminer git compile in linux ( sgminer 5.1.1-17-g666 ) and ALL of them show share rejects on a regular basis mining on quark.usa.nicehash.com:3345 ... note - this is mining 'just' quark ... not algo switching ... i have seen sgminer ( git ) show regular share rejects switching algos also - though the testing is no where near as comprehensive as what we have already done on each algo singularly ... take a look at the current stats on the test bench as we chat - https://www.westhash.com/?p=miners&a=12&addr=15umzHXF8NzXA4FywmeFbrDHgL8WcPs3wx ... we are working alongside another long term respected user here on bct ( who we wont name for the moment ) on an algo switching script for linux using nicehash and ( originally ) yaamp ... obviously nicehash is the main testbed now until a new yaamp comes about ( which is currently planned ) - but these tests show proof that the farm in all its different incarnations and hardware / software makeup will never submit shares that are ALWAYS accepted ... this is not to single out any one pool or another ... it is to show ( at least to us ) that rejected shares can only be minimized - not altogether eradicated ... no matter how 'good' the mining software - or mining hardware - or stratum software / pool software is ... so even though you may not be aware of the share rejects - our mining proves otherwise ( at least for the miners we have tested internally ) ... we are putting together all the donation links through nicehash - which means that even though share rejects happen - it is almost negligible in the long run ... btw - havent heard back from support whether the other ip addresses we have submitted ( for the donation links ) have been whitelisted yet ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) ... #crysx Strange, I'm using the Russian Miner at westhash and I have zero rejects over a long enough period of time to be significant. . Literally zero, I've never seen anything like it. I'm much closer to the server than you are, maybe it's the latency that's causing the rejects. that could be ... latency plays a part - though i would think not a major one ... i could be completely wrong to that end ... please dont take what i have written as law ... the testing process is not clinical nor documented and is far from perfect ... BUT - it is enough to show us that share rejects DO happen ... no matter how small ... AND that the majority of miners we have running are gigabyte 750ti oc lp cards running ccminer-spmod letest git compile using cuda 6.5 in fedora 20 x64 ... so there are bound to be discrepancies ALL OVER the place ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) ... especially when using different hardware on different platforms with different software ... we are setting up various avenues to test the miners and software combinations ( direct to pool - via stratum proxies - via different pools like mintsy for example ) - but we wont go into too much detail with them ... as said earlier - the rejections are not that much to worry about ... especially when the full farm is mining the one algo ... as mentioned earlier - this is something that we are working in tandem with the other party ( who is the coder ) to algo-surf and minimize rejects ... one thing is for absolute certain though ... the amount of time that goes into the setup and compile and stuffing around with windows is beyond what we are willing to partake as being acceptable ... the majority of our farm is linux ... even with the shortfall of support for the os - it is by far the superior of the platforms in terms of efficiency - setup time and ease - and downright stability ... my personal experience over the last few days have been a hair pulling - time wasting - wanting to punch windows in the face - kind of situation ... never again ... unless microsoft pull a rabbit out their backsides with windows 10 - our farm will be exclusively linux based ( as it has been ) ... period! ... but that is another story in itself ... #crysx
|
|
|
|
chrysophylax
Legendary
Offline
Activity: 2870
Merit: 1091
--- ChainWorks Industries ---
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 03:49:07 PM |
|
i mean yaamp's code, not site, just try and see for yourself, multialgo switching makes huge difference
i see what you mean ... on this fact alone - it does seem that way ... multialgo switching has abrupt changes to contend with and so suffer share rejects everytime the switch happens ... this is a loss of shares and productivity - even though they are small between switches ... we have seen this happen with our own miners / farm ... but that is with ANY multialgo multipool - not just yaamp.com ... look at epsylon3 .. he is one hell of a developer - and now running his own yaamp-coded pool with more algos added ... it sort of goes to show that if even a seasoned developer like him is willing to look at - setup - and improve the code - then the pool code is still on a reasonably good level ... nicehash has the same sort of issues ( though they dont mine coins - just divert hashrate ) with multialgo switching ... so much so - that they devote an entire section on the website to it ... single algo / single coin pools may be more 'efficient' ( for lack of a better word ) at mining - but we believe that the losses are so minimal thats its close to negligible ... #crysx I don't have a problem with rejects upon switching algos on any multipool. I don't use the pools' multialgo features, but I also incur the overhead of killing and restarting ccminer. I don't think there is a truly elegant way to switch algos. The pool signals by disconnecting so any unsubmitted shares are lost. Whether the miner reports that as a reject is irrelevant, it's lost work. To have a seamless algo switch requires a cooperative disconnect where the miner is allowed to complete the work in progress. agreed ... which also means that the miner and stratum need to work hand-in-hand with one another to become as seamless as possible ... which lends itself to have the miner AND the stratum backend written by the same dev ... or at least devs that work WITH one another on the same level ... wolf? ... ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) ... #crysx
|
|
|
|
NiceHashSupport
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 04:47:59 PM |
|
We had plans to extend stratum with support for switching algos without connect interruption, but it turned out that disconnects are very tiny issue when miner is switching algo. The most time is taken for switching gpu kernel. Thus we decided not to persuade this since miners wouldn't benefit at all, but rather made miners switch algos more rarely. There are no issues with multialgo switching for about a year now. If you have them, you should really check out mining software or proxy software you use, because there is something seriously wrong.
|
|
|
|
chrysophylax
Legendary
Offline
Activity: 2870
Merit: 1091
--- ChainWorks Industries ---
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 04:56:21 PM |
|
We had plans to extend stratum with support for switching algos without connect interruption, but it turned out that disconnects are very tiny issue when miner is switching algo. The most time is taken for switching gpu kernel. Thus we decided not to persuade this since miners wouldn't benefit at all, but rather made miners switch algos more rarely. There are no issues with multialgo switching for about a year now. If you have them, you should really check out mining software or proxy software you use, because there is something seriously wrong.
hence the rebuild of the farm - the software - and the systems we have in place ... ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) ... #crysx
|
|
|
|
|
ffpool.net
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 82
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 07:46:22 PM |
|
Thanks for your feedback, good to hear ! Our Quark Hashrate has also quadrupled over the last hours, we're over 600 MH/s now constantly and are getting a lot of lucky Sharkcoin blocks. Even more than 1 block per hour ;-)
|
|
|
|
Epsylon3
Legendary
Offline
Activity: 1484
Merit: 1082
ccminer/cpuminer developer
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 07:52:11 PM |
|
my yiimp will not do that... auto sell coins to Btcs... i don't like this system... which auto kill the coins value
|
|
|
|
joblo
Legendary
Offline
Activity: 1470
Merit: 1114
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 09:03:10 PM |
|
|
|
|
|
joblo
Legendary
Offline
Activity: 1470
Merit: 1114
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 09:12:23 PM |
|
Profit estimates on yiimp seem off. For example zr5 in the pool status shows an estimate of .0001 for the current & 24 hr estimate but .7772 for the 24 hr actual and in the pool details it shows ziftr paying .0014. Also the block reward in the pool details is different from the block list. The other clones seem to have got it right.
I also noticed that yiimp mines two algos not supported in Epsylon3's fork of ccminer. I presume they will be added shortly as there appear to be open source versions of ccminer for both floating around.
|
|
|
|
Epsylon3
Legendary
Offline
Activity: 1484
Merit: 1082
ccminer/cpuminer developer
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 05, 2015, 10:22:12 PM |
|
yep, i need to read the code about these estimations... doesnt look good :p
drop algo is not on my ccminer (its nscrypto one, the dropcoin dev), c11 is supported on my linux branch... didnt released the v1.6.6 yet
|
|
|
|
chrysophylax
Legendary
Offline
Activity: 2870
Merit: 1091
--- ChainWorks Industries ---
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 06, 2015, 05:36:12 AM |
|
We had plans to extend stratum with support for switching algos without connect interruption, but it turned out that disconnects are very tiny issue when miner is switching algo. The most time is taken for switching gpu kernel. Thus we decided not to persuade this since miners wouldn't benefit at all, but rather made miners switch algos more rarely. There are no issues with multialgo switching for about a year now. If you have them, you should really check out mining software or proxy software you use, because there is something seriously wrong.
hence the rebuild of the farm - the software - and the systems we have in place ... ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) ... #crysx just changed over to the eu stratum for testing over the next few days ... we are using the exact same setup as the usa stratum - but just diverted the hashflow ... https://www.nicehash.com/?p=miners&a=12&addr=15umzHXF8NzXA4FywmeFbrDHgL8WcPs3wx ... will keep an eye on the share rejections ... will also be testing the current pools that are active with yaamp code soon via quark with the same settings ... though that will probably happen next week ... too much to do in the backend at the moment ... #crysx
|
|
|
|
chrysophylax
Legendary
Offline
Activity: 2870
Merit: 1091
--- ChainWorks Industries ---
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 06, 2015, 05:39:03 AM |
|
yep, i need to read the code about these estimations... doesnt look good :p
drop algo is not on my ccminer (its nscrypto one, the dropcoin dev), c11 is supported on my linux branch... didnt released the v1.6.6 yet
looking forward to the changes mate ... #crysx
|
|
|
|
GingerAle
Legendary
Offline
Activity: 1260
Merit: 1008
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 08, 2015, 10:12:41 AM |
|
Any chance cryptonight could get into this multiago pool? I think tsivs ccminer code would need to be plopped into the regular ccminer code (he eventually stripped all algos except cryptonight from his fork.
|
|
|
|
joblo
Legendary
Offline
Activity: 1470
Merit: 1114
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 08, 2015, 01:56:48 PM |
|
Any chance cryptonight could get into this multiago pool? I think tsivs ccminer code would need to be plopped into the regular ccminer code (he eventually stripped all algos except cryptonight from his fork.
ffpool and hashpower each have their own thread now so you might be better off posting your request there. yiimp is a test pool so probably isn't suitable. The biggest ccminer forks are produced by Epsylon3 and SP_. They also each have their own thread tracking their development. I think it's a great idea to get cryptonight in a yaamp clone. I'm not aware of any existing cryptonight pools that auto exchange.
|
|
|
|
Epsylon3
Legendary
Offline
Activity: 1484
Merit: 1082
ccminer/cpuminer developer
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 08, 2015, 02:13:36 PM |
|
i will do a thread for yiimp, but it will never do auto exchange... only mining ccminer/cpuminer compatible coins
I was tired to search pools for some algos, so having our own one to do tests is very useful. I'm adding algos one by one and check if all is ok...
Actually yaamp doesnt seems ready at all to mine "multialgos" coins... its one of my current tests...
|
|
|
|
yaamp (OP)
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 11, 2015, 03:34:10 PM |
|
It is highly recommended to start from this fork if anyone has the idea to do anything with the yaamp code. https://github.com/tpruvot/yiimp
|
yaamp.com
|
|
|
hoosen
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 163
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 12, 2015, 03:23:50 PM |
|
Yet Another YAAMP based multipool. Not that many coins for now, but we'll add more soon. http://hashpower.coNo scrypt coins to mine?
|
|
|
|
Epsylon3
Legendary
Offline
Activity: 1484
Merit: 1082
ccminer/cpuminer developer
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
July 13, 2015, 02:14:40 AM |
|
Tx ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) I have at least 50 local commits here to submit later... some for fedora, some new algos, some new "dangerous" yiic commands (distclean deleteuser) and new exchanges/markets apis (alcurex,banx and cryptopia)
|
|
|
|
|