Topic: [NemosMiner] multi algo profit switching NVIDIA/CPU miner - page 77. (Read 289464 times)

Same to me - logging Ahashpool. I am trying to run manulay. BTC is OK, but second grab not.
Something I'm doing wrong, I think.

Guys is yiimplogger 2.1 working for you?

Mine only will update the BTC price and thats it...The style of the logger is also not the same as the demo :/

Not if their business model is occasionally 'get hacked' and profit from those 'losses'  

I'd be interested to see the results.  If I can find 4 more 1070's at a price that doesn't violate me I'll run some more comparisons too.  

Yeah, NH can bring such a huge amount of hashpower to bear on any algo, so if the coin that is the "flavor of the minute" has a tiny float (er, circulating supply) then difficulty skyrockets, locking out the small miners while NH simultaneously shoots itself in the foot and vacuums all the coins. Their business model is kind of nutty, actually.

As for hsrminer, almost every time I've run it - note that "almost" - it starts off roaring then mysteriously drops by a certain percentage after awhile. Usually the hashrate drops after 1-3 hours, but there have been times where the hashrate was stable for more than 6 hours, so I'm not sure what to think about that.

However, in every single case so far the actual earnings have been substantially lower than the hashrate would otherwise have me believe; in my latest test I earned 76% of the coins that hsrminer said I would after 8 hours of run time on a pool that finds a block every 4 minutes (so a short run time is less likely to skew the earnings). That would make hsrminer slower than its closest competitor, ccminer-klaust 8.19, but I will need to run the latter on the same pool for the same length of time to confirm that.

As for NH's benchmarking, I can't say for sure whether it is based on client-side reported hashrate, or server-side reported shares... The latter would be more accurate but people might complain about getting shortchanged enough to make NH not use it.

Ah, gotcha. The 2 Neoscrypt coins I have been mining directly are Feathercoin and Trezarcoin, but FTC is only rarely worth mining, I've noticed (based on difficulty vs. price); if FTC is the only choice of Neoscrypt coin on MPH then I can definitely see your point.

So it looks like I need to do comparison between whichever pool wins this round and Ahashpool, then.

Yeah that's quite the circus alright. From what I've observed, NH is very prone to price manipulation for a number of algos. Due to this, I've decided to just limit algos for both AHP and NH, in hopes of finding a combination that isn't going to result in a lot of downtime (neoscrypt for example takes quite a while to "warm up").

Thoughts on hsrminer's (and therefore, Nemo's) numbers for neoscrypt? Some people are saying it's pumped up, but I'm not mining neo on AHP, and NH doesn't have precise hash numbers outside of benchmarking. Actually, as I write this, I'm wondering if benchmarking on NH is client or server-side. With people saying hsrminer falls off severely in hash after 15-60mins, I'm wondering if that was to pump up benchmark numbers before actually mining.

It has more to do with the number of coins and what coins are mined per each algo.  MPH only has feathercoin on neoscrypt while Ahash has 8 other coins in that algo.  Equihash might be more profitable on MPH but while mining on Zpool I've rarely had my miners turn to that algo.  Also MPH doesn't really have a way to mine xvg which even though it's down is still an anchor for blake2s and lyra2v2 and x17, all of which can be really profitable. 

I've done tests between zpool and ahashpool and if you are using the profitability switching ahash killed zpool by almost 20%, but I'd be really interested to see a run up of MPH or Nicehack against ahash.  I was doing one with ahash against hashrefinery but they shit the bed and mined the wrong side of the xvg fork and it screwed everything up.

FWIW I use https://ethmonitoring.com to keep an eye on my rigs running nemos. That isn't an affiliate link just sharing a tool I find useful.

I found the temperature alerts catch a hung GPU well e.g. alert me if a GPU temp is below 35˚C for longer than 3 minutes usually means it's sitting there doing nothing. Then you can restart the rig from the app too.

Some of the alerts don't work so well because nemos is switching algos e.g. hashrate related ones because they vary so much.

I also have alerts for GPU temp over 70˚C (it's a hot summer downunder!) and if a rig is offline longer than 5 minutes and 10 minutes and 30 minutes –
 just to make sure I get the message!

One rig is free to test it but if you do make sure you get the 'monitoring' app to use with nemos – the 'control' app is an all-in-one which includes miners and some other stuff.

Nope, haven't tried Ahashpool. I've looked at the website, but that's about it. I'm curious as to why you think MPH sucks for Nvidia - Nemosminer seems to use Neoscrypt, Equihash and Skein the most and the GTX 1080 does well on all of those algos - or are you saying that Cryptonight, Groestl/Myriad-Groestl coins are the most profitable on MPH (but wouldn't that change on a minute-by-minute basis?)?

MPH vs NiceHash isn't really a good comparison.  Most of the most profitable algo/coins for nvidia aren't really on MPH.  Did you try Ahash?

Yeah, I have to say I feel exceptionally dirty even testing NH out, but I'm doing it in the name of science. I know NH can totally wreck new coins by pumping up difficulty so high your earnings get cut by 90% or more - I saw that with DERO, for example - but I feel I have to test them out to make sure I am not leaving the proverbial turn unstoned. Or something like that.

I did figure that comparing the two auto-switching miners concurrently for 1 week would be fair overall, but if not I'm not sure I want to tie up 2 rigs for more than 1 week, anyway. At some point these things need to start earning their keep, after all. 

I've got quite a circus going on right now, actually, with 4 separate rigs doing different things at the same time. The relevant comparison here is between Nemosminer on MiningPoolHub vs. NiceHash Miner straight out of the box, so to speak. Both Nemosminer/MiningPoolHub and NiceHash are auto-switching between algorithms - and not always choosing the same one, I might add - and each get their own GTX 1080 to use (running on separate computers).

I also have a 3rd rig running a Neoscrypt miner (Gateless Gate - AMD only) on Zergpool which auto-switches between multiple coins of the same algorithm and converts all proceeds to BTC, just like Nemosminer and NiceHash. I am mainly testing Zergpool out of curiosity, but am prepared to be surprised should it show up the multi-algo auto-switching miners.

Don't forget NH will be actively looking at new shitcoins to mine (rape) to make an extra buck. Pools like Miningpoolhub are pretty static with the coins they switch between... so if their coins aren't having a good streak, your profitability on mph will be much more likely to dive.

Excellent, are you running the tests concurrently on similar rigs or consecutively? I am also doing a small test of Zergpool which has the ability to auto-switch between multiple coins using the same algo (I'm testing Neoscrypt). I don't expect that to be the most profitable use of mining hardware, but if it is better than mining a single coin then I might consider it. Regardless, whichever one earns the most will next get tested on my real mining rig of 6x GTX 1060 for 1 week.

As for the BTC fees, I suspect they will be higher with NM/MPH because I have a verified account at Coinbase, so a transfer from NH should be free... I hope...

That said, I actually don't know what to expect for fees when I transfer BTC from MPH to either my Bittrex trading account of my Coinbase useless-for-anything-except-fiat-conversions account.

One potential issue with Nemosminer is the over-payment of devfees with miners that do their devfee mining first if the most profitable algo changes frequently (which it does, from what I've noticed). This argues for using a much longer interval between profitability checks, but increasing that time defeats the purpose of an auto-switching miner. Finding an optimum will be tough, I suspect.

Haven't gotten to MPH testing yet, but I'm running Nemos -> ahashpool24h currently and approaching the 24h mark in about 7 hours. It doesn't look like I'm going to be able to produce the estimated results from Nemos, but we'll see soon enough.

I'm guessing you're running hsrminer -> NHML (saw you in the thread) for Neoscrypt? Or did you just point your Nemos -> NH and forgo the NHML GUI? Either way, looking forward to your findings.

@nemo: I have several rigs with 6 or more GPUs.
So if for any unspecific reason one of the GPUs stops, the whole rig stops, and a lot of money will be lost until I can restart the miner in the evening.

Will there be a "-SelGPUCC 'ALL'"    instead of    "-SelGPUCC '0,1,2,3,4,5'" feature some time? This would be very helpful.

Ok but I read on this topic 24h is too short and suggested to mine  at least 1 week for an estimate profit 

Thanks.
I'm running the same test but over a week. Going to add awesome miner to the comparison also.
Then after thats sorted out; will try out various pools.

Im curious to see how the larger NH fees will change your final BTC amount.

Day 3 of my comparison between Nemosminer on MiningPoolHub vs. NiceHash just ticked over and so far NiceHash is well ahead with 1.08846 mBTC earned vs. 0.89968 mBTC for Nemosminer/MiningPoolHub, or about 20% more.

I'll do a more detailed write-up in a new thread when the test ends on Thursday (after 1 full week of mining) but NiceHash has been ahead pretty much the entire time so it's going to take a miracle for Nemosminer to win at this point.