Topic: Team Black Miner (ETHW ETC Vertcoin Ravencoin Zilliqa +dual +tripple mining ) - page 49. (Read 35090 times)

That's your problem right there. 2150 is not stable on all cards, and it clearly shows that in your invalid share rate.

Unless you've pushed over 50k shares or the statistical hypothesis being tested has a very large difference, the randomness intrinsic to mining makes any claims of faster or slower mining rates meaningless.

They are the same and considered normal with every other miner except here I am at 1300 Mhz core clock and not 1200 Mhz core clock otherwise one of my cards has a floating hash rate,goes to 47,50,55 and 60 max with 1200 Mhz core clock,with 1300 Mhz it is stable.The memory clock everyone use it to the max at 2150 and that is my number also.Maybe the problem is the pool ezil.me which is not a huge one and probably has something to do with this as my settings are the perfect ones for everyone else so why shouldn't they work for me.



That average in the pool is what matters more to me as is based on that that the pool pays and the highest is with this miner so far,maybe just luck but I don't know.I never had an average of 280 Mhsh from 289 Mhsh max power with any other miner in the pool and it is like that from quite some time now.That explains why the payouts have been better than with any other miner,because of the stability.

Thanks for the link, had already switched away due to multiple missing features and antifeatures such as built in OC and intentionally blocking proxies, but this seals the deal.

I wasn't ever able to achieve anything close to the reported hashrates either, even with the "high share difficulty and using a pool with stales" settings (around 2 rigs, each putting out around 10k shares of 8G diff on 2miners). Overall hashrates were reported 3% higher but showing as 4-5% lower than reported poolside. This by itself is not statistically significant evidence, but combined with the data above it's pretty strong proof.

There is probably no evidence of outright malice, but the claim that the miner will only be able to reach its reported speed on a configuration that actively impedes statistical testing feels too convenient.

To the dev, I hope you figure out the reason behind the low pool share reporting, because if not this miner will end up with phoenixminer's reputation. It's not impossible to develop a new miner with actually faster hashrates, TRM did it, but if you can't back it up with evidence then it's all for naught.

---

To be clear, these are not 2% stales, these are 2% invalid shares. Your overclock settings are way too high, they are crashing because they are unstable overclocks. Dial them down to a more sane number.

My stable OC 6800(xt) show 67-68M on TBM.

Also, SAM or BAR do not help hashrate. They increase the rate data can be pushed to the card, but that is not a bottleneck for mining, if it were, x1 risers would not be used.

[moderator's note: consecutive posts merged]

Perhaps the Vulkan API can be used on NVIDIA LHR cards to do a 100% unlock..

I can accept that as long as the Team Black Miner is the only one out of all other miners who has not crashed yet for more than one day.Only Phoenix miner ran for me for just 3.5 days one time and every time after it crashes after several hours or maximum 1 or 2 days.So far I am seeing the best stability from Team Black Miner and I see the hash rate to the pool is like 153-155 Mhsh from 173 Mhsh reported by the miner which I am happy with,it is consistent different from Phoenix miner which was not consistent at all in the pool.So far I only pray it will not crash like all other miners,as for the deviance I can accept that.I made more ETH+ZIL during this day which I mined with Team Black Miner than with every other miner because every other just crashed and it was not consistent.I hope this does not crash and I am happy.I know how mining works but what is important to me is how much I make in the end of the day,week and month.Yesterday before I started mining with this miner I had an amount of 0.015 ETH in the pool,now which is just 1 day and 1 hour more I have 0.02xxx ETH meaning it is running better in the long run as long as it does not crash,the average will be more stable.
It has a few more stale shares than other miners but I am sure it will be fixed in other versions as I have a 100 Mbps fiber connection and with all other miners I had like 0.75% stale,now I have near 2% but I can accept this also as long as it is stable and does not crash.

I took this long to write because I am way to excited seeing a miner not crashing with my Powercolor RX 6800 XT-s,every other one has crashed so far.

The test was performed in a simulator program on utra low difficulty.

To get the best results with the Team Black Miner, mine with high --xintensity on a high difficulty pool that support stale shares.

For more info

https://github.com/sp-hash/TeamBlackMiner/blob/main/FAQ.md

Honestly I'm not surprised about those results.
Happy to switched back to GMiner.

You should stop mining if you still haven't figured out how mining works in years ...

What gives your performance, stats as always on 1 million shares?:

even PhoenixMiner is rated higher! lol

https://www.hashrate.no/tests/ampere

I'm curious if there's any plans on dropping OpenCL on modern AMD chips in favour of the new Vulkan API (which contrary to popular opinion isn't a replacement only for OpenGL, but for both OpenGL and OpenCL).  I hear that the Vulkan libraries are 30-60% faster than OpenCL (in the linux 5.16 kernel (amdgpu + AMDVLK / vulkan-radeon) anyway).

For the moment I am to excited about that 67 Mhsh seeing in the miner and I am staring at the Anydesk from almost half hour now.I don't even have a new motherboard with BAR option to give another boost otherwise I believe I would be seeing 75 Mhsh without problems here.I will update in due time to the new version,no worries about that.

Hopefully it doesn't crash after sometime like all other miners.

Note that the dag cache isn't implemented in 1.27 for amd cards, so when you mine eth+zil you loose 2 dag generations every 90 minutes. A few percent. This is fixed in v1.28

I have power limit at 71 and I don't intend to change it because of electricity price not being that cheap here and I am happy with that 42 Mhsh,all other miners had 42.5 Mhsh but who cares for 0.5 Mhsh lost when I gained about 8 Mhsh in total as in the previous pic it is just a time when it was near 65,these cards 6800 XT-s with your miner (always if it does not lie about the hashrate ) are running near 67 Mhsh stable wow.

In epoc 455 the 1080ti need abit more power and coreclock to keep the hashrate. In msi afterburner increase the powerlimit(80 ?) and the templimit (85), and try to get the coreclock stable at 2050... 47 MHASH

Here with --kernel 0 --xintensity 1024...

core: 2060
mem: 5700

Lower kernel numbers are stable on higher clocks.

It is running well from a few minutes without the kernel 0 option.Your miner has given me 6-7 Mhsh in total more in the RX 6800 XT cards,the strange thing though is I don't see any improvement at all at Nvidia where I was expecting a big gain,it runs at near 42 Mhsh with the pill ongoing in the background.Anyway if this miner does not crash like the others is going to be my final decision to go with this.

This is not a crash. Do you see the white square in the window? Rightclick with the mouse in the window. And the output will resume. A problem with windows 10. You can try without --kernel 0 in the bat. Then the program will autotune. It might go faster.

Now it hangs here for several minutes after not doing anything,this same error happens with lolminer only they got stuck.All other miners crash after a few hours except Phoenixminer which runs a few days and then it crashes also.I maybe have an AMD card with some form of defect but don't know,they are using Xintensity as dynamic,maybe I can add another value to this to try.

There was an error in the docs. The correct parameter is --worker-name and not --worker_name

so rename to --worker-name

I did it like that exactly.Here is a picture of the error anyway and of the bat file configuration.Tell me what is wrong.








Edit: it worked after as suggested changed worker_name to worker-name so thank you.

https://github.com/sp-hash/TeamBlackMiner/blob/main/HOW-TO.md


The correct parameter in v1.27 is

--worker-name

Same error

--worker_name command not found.