Topic: [NemosMiner] multi algo profit switching NVIDIA/CPU miner - page 74. (Read 289464 times)

Hmm... the pool I was using for the Neoscrypt miner test finds a block every 2-4 minutes on average so that shouldn't result in more than about +/- 0.5% difference in earned coins over a 24 hour period. I could redo the test mining in parallel with a single GTX 1080 for each - as I am doing right now comparing dstm to bminer - but I seriously doubt that will erase the 20 percentage point difference I got between ccminer and hsrminer.

I should also note that what inspired me to do this test was that I switched my one real mining rig (6x GTX 1060) over from ZEN to TZC for a few hours because ZEN difficulty was straight up nuts (basically tripled) and I wanted a few more TZC for staking. Since hsrminer appeared to be faster that is what I used and I let it run for nearly 5 hours. After that time frame - while keeping an eye on the current difficulty on minethecoin.com - I found I had only earned about 75% of what I should have, so I decided to run ccminer klaust for 24 hours on the same pool then run hsrminer again, except this time for 24 hours; first hsrminer run (~5h) gave me 75% of expected earnings while second run (24h) gave me 70% of expected earnings. Klaust, sandwiched in between the two hsrminer runs, gave me 90% of expected earnings.

Pool luck could still have played a role in the discrepancies. You really do need to run in parallel to have a near 100% accurate comparison. I do see my hashrate on the pool being quite a bit higher when I use hsrminer vs ccminer KlausT. Might be worth doing that test again.

Meanwhile, the ahashpool situation is concerning.

Yes, I ran two rigs at the same time for the MPH vs. NH test. Each was a single GTX 1080.

The miner software test was run consecutively, but I normalized earnings by using the average difficulty for each 24 hour period for each miner so they were evaluated based on the coins earned vs. their reported hashrate vs. the average difficulty for their 24 hour period.

What is not really clear to me, you ran 2 mining rigs at the same time? Or did you test them after each other. Because market is dropping day after day now so could also make a difference in earnings.

While comparing Nemosminer/MiningPoolHub vs. NiceHash I was also doing a Neoscrypt miner comparison between hsrminer and ccminer KlausT (8.19) and I think I may have found one, if not the, reason why Nemosminer/MPH lost to NH: hsrminer lies like a rug about its hashrate.

I tested the two miners by doing consecutive 24 hour runs with a 6x GTX 1060 rig on the official Trezarcoin pool because it is one that I trust and it finds a block every 4 minutes or less on average, minimizing the effect pool luck might have on the results.

After 24 hours I stopped the miner-under-test and looked at the average difficulty reported by minethecoin.com for the past 24 hours and then iterated the hashrate until the number of coins on minethecoin matched what I actually earned. If there are errors in minethecoin's calculations, then they would likely affect both miners equally.

The results were not pretty. For hsrminer, anyway.

ccminer klaust reported an average hashrate of 4.1 MH/s and should have earned 123 coins; actual earnings were 111. Effective hashrate was 3.7 MH/s.

hsrminer reported an average hashrate of 4.3 MH/s and should have earned 126 coins, but actual earnings were 92!?! Effective hashrate was 3 MH/s!

In other words, hsrminer was 30% slower than it claimed, and nearly 20% slower than ccminer klaust. Needless to say, this radically changes the profitability matrix generated by Nemosminer's benchmarking script and so I would like to redo the Nemosminer/MPH vs. NH comparison except this time with ccminer klaust (or one of the existing ccminer forks, at least) for the Neoscrypt miner.

EDIT - I am not changing any of the above, but notifying any who come across this that I did another test of ccminer and hsrminer, except run concurrently mining to different addresses on the same pool, and this time hsrminer was 11% faster (though claiming to be 22% faster).

Hello gyus,

i downloaded nemos miner (latest) and i ran all benchmarks (i added phoenix miner for ethash)
when the benchmark reach phoenix miner it just dosent stop and keep mining and never run other benchmarks or switch to other algo.

anyone faced the same pb ?

remember calling me a noob a week ago?

what is the most profitable pool?

So the afterEFX of spreading ahashpool all over the place have finally hit the pool rewards.
I wonder wich one you'll flood next 

FWIW, I am informally testing both ZergPool and HashRefinery on "mine any Neoscrypt coin, convert to BTC" mode and both are doing quite well. I will likely do a formal comparison of them next week, but I can already tell you that I have earned significantly more on ZergPool than I would have mining a single coin like TZC for the same time period. Their servers are in Latvia so the ping time is a bit long for me here in the US - about 150ms - but I am averaging about 0.1% rejected shares despite that.

Thanks for doing this.  Always good to see what the different options are.

Testing did stop about 3 hours ago and the total balance at NiceHash is 0.00207672 BTC, with 0.00126187 BTC sitting on their wallet, ready to send to me, and another 0.00081485 BTC tied up on my miner - I really don't get how that even happens but, whatever... I just have to run NH a little longer to get that balance up to 0.001 so it transfers to my (what? NH wallet?).

Meanwhile, MiningPoolHub is sitting at 0.00180649 BTC but there are still balances in several coins that are in the process of getting converted to BTC and who knows how long that will take to complete. Given that it has already been 3 hours I doubt they are going to get processed and confirmed anytime soon so maybe tomorrow morning I'll have the final tally for MPH. That said, it doesn't look like the balance will increase all that much so as of now MPH got trounced by NH.

EDIT - forgot to specify where the balance from NH miner goes... probably because I don't know why it has to go in the first place.  

UPDATE - Balance on MPH has climbed to 0.00187639 BTC over the last few hours and there is still one coin that needs to be exchanged.

Update 2 - Final balance on MPH came to 0.00188193 BTC, or 10.4% less than NH for 1 week of mining with a single GTX 1080.

That would make some sense; perhaps I'm tinfoil-hatting too hard.

But since these 3 days have been red, shouldn't the 24hr average have adjusted by the second day? I was missing the mark on the 48hrs following my initial 24hrs.

It has a lot to do with how the pools report their profit.  Right now most coins are dropping in value faster than they can be mined.  If you look at any of the exchanges all you've seen for the past few days is a sea of red.  So if something like Nemos is going off the 24h reported profit from a pool like ahash it is showing a higher profit because a day ago it was better than it is right now XVG for example was trading at 800sats a week ago and the current price is 650.  That's a pretty big drop over the week and it's a linear drop instead of just a fall off which is why the reported profits are so far off.

I just hope that the crappy profits calm the GPU market enough so I can pick up some more cards and build a couple more rigs.    I'm also curious if Nvidia has some super secret plans to flood the market and kill all the resalers.

It's certainly not just you guys. I've been on AHP for the past 3 days, and not a single day did it come to anywhere close to the expected earnings quoted in NEMOS. Now, I know, NEMOS certainly isn't the most accurate measure of profit, but I don't see an excuse to be ~20% off projections. I've even looked at the numbers from

http://pools.you-stupid-woman.com/

and they're nowhere close. I thought I must've been doing something wrong, but now that it seems I'm not an isolated case, I'm pretty sure something's up behind the scenes.

https://bitcointalksearch.org/topic/yiimp-pools-zpoolahashpoolhashrefinery-making-large-payments-to-non-miners-2835350

That thread certainly isn't exactly reassuring either.

If anyone's interested, I'm running 7x1070Tis and 1x1080Ti. Was quoted to make ~3.6 mBTC by NEMOS, and only received ~3.0 mBTC. I had about 20-30 mins of downtime in that 24hr time frame, but that certainly shouldn't diminish profits significantly. In the other 48hrs that I tested, I was also missing about 20% if not more, but due to increased downtime and sloppiness, my case for those is weaker (but still valid IMO).

Currently I'm on NH, and the projections seem relatively accurate. I'm expected to make ~3.2 mBTC (it's a slaughter out there right now eh) by the next 8.25hrs, so I'll report back then.

http://pool.hashrefinery.com/
They still don't have any xvg algos right now.  Nothing since they screwed up the hard fork.

Is there any miner like neosminer that has auto switch function for nvidia cards under linux?

After 3 hours it seems Zpool Blake2s (XVG) mining with C=XVG could exceed unimining

unimining blake2s - 0.05553 (24hr actual according to unimining)
zpool blake2s - 0.07389 (24hr actual according to zpool)

Can't get hashrefinery website to come up.

Zpool for the most part is NOT selling/buying since 95% of what is mined is already XVG, so I don't have to worry about buying on ask/selling at dump pricing.

I was killing it at hash refinery until they did the same thing and then disabled all xvg algos.  I felt like they did better exchange rates there.  Zpool when I mine to a different coin other than btc seems like they calculate selling it at dump pricing and buying at ask pricing into it.  I could be wrong but the numbers weren't adding up when I was doing that.

Was using UNIMINING very nicely up until this morning when they started to mine against a bad fork!  They haven't fixed it so using Zpool with c=XVG right now so that I get paid in XVG