Quote from: jch9678 on July 05, 2015, 03:25:09 PM

Quote from: chrysophylax on July 05, 2015, 03:15:24 PM

Quote from: NiceHashSupport on July 05, 2015, 02:15:59 PM

Quote from: chrysophylax on July 05, 2015, 01:52:14 PM

Quote from: unklo on July 05, 2015, 01:25:04 PM

Quote from: chrysophylax on July 05, 2015, 01:13:30 PM

Quote from: unklo on July 05, 2015, 01:05:35 PM

Quote from: pallas on July 05, 2015, 07:42:44 AM

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

...

#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

... 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