ok. we will fix it in v1.24 https://github.com/sp-hash/TeamBlackMiner/issues/110
0m [2021-11-08 01:18:36.381] GPU0 speed 0:98.59 1:99.81 2:99.97 3:100.00 4:99.98 5:99.96 6:99.98 7:99.97 8:100.00 0m [2021-11-08 01:18:36.381] GPU0 using gpu kernel 3 0m [2021-11-08 01:18:36.912] GPU1 speed 0:99.44 1:99.66 2:99.98 3:100.00 4:99.97 5:100.00 6:99.99 7:99.98 8:99.99 0m [2021-11-08 01:18:36.912] GPU1 using gpu kernel 5
The two kernels are running at the same speed. You can try to restart the program a couple of times to see wich one is the best, and then run with --kernel x to skip the autotune. [moderator's note: consecutive posts merged]
|
|
|
How to get rid of this Red Error shown in screenshot attached!
Upgrade to v1.23
|
|
|
Nice Work! can you show diff. at found shares? & show correct pool difficulty? Here 8.72G Shouldn't be too hard to fix. The difficulty is different for the ethproxy protocol than ethereum stratum v1.0 so the output in the stats depends on the pool.
i dont really want to switch to ezil.me , since shardpool allows me to keep mining eth on ethermine p.s. got same issue on 1.23 also , thats why i made post , so this is something recurring.
try without --xintensity -1. What is your complete bat file ? so we can debug [moderator's note: consecutive posts merged]
|
|
|
runing with : --tweak [0,1] --power-limit [0,70] --dagintensity [9,1] --xintensity [1024,1024] running more stable for me ...still trying to find what number for --tweak. Is it the same for -mt as we know for pascal only or something else?
Tweak and memtweak is not the same. On the gtx 1080 you should run with the pill in the backround. tweak can be combined with xintensity to improve hashrate. High tweak together with high xintensity. Or high tweak together with low kernel number and high intensity. On low intensity use without the tweak parameter (0) or set it to 1.
|
|
|
at some point shares stops validating on cpu , experiencing that on 2 separate boxes, one with 3080 & 5900x on shardpool eth/zil , second with 2060 cards and 3700x on ethermine.
Did you try on ezil.me? Try v1.23, there are some bugs fixed. We will put some cards to investigate and try to reproduce.
|
|
|
This has been running for 7 days so it should have averaged out any hashrate fluctations, on TBM 1.22 kernel 8 default xintensity.
You get best results if you change the pool, but you can try 1.23 with --xintensity 144. It's not only about stale shares, it's the way the protocol on the pool works and how much work the miner needs to throw away. Jobs are too frequent. v1.23 has an increased performance on lower --xintensity so should work better on the hiveon pool.
|
|
|
always the same same with --dagintensity 1
Upgrade to v1.23. If it doesn't work Lower the memclock -100, -200 until it's stable.
|
|
|
Hey sp_, Thanks for your work, - just sharing some feedback: 5 x RX 6600 XT rig went from 160 Mhs to 171 Mhs. Is that a solid 6.8% increase or what? Further, another rig of 8 x RX 5600 XT's went from 336 Mhs to 350 Mjs, however, - after swapping to TBM (from TeamRedMiner), GPU 3 started "dropping". It stops mining and nothing happens, - other cards remain online. I dropped the memory clock by 15, then GPU 0 started doing the same - a "gap in hashrate" can be visitble in the below snippet. So, I ended up decreasing the memory clock by 15 as well. However, an increase of 4.1% is very, very pleasant to see. Thanks again, all the best, Ric. Nice results. If you want to remove stale shares on ethermine use --xintensity 24 or perhaps --xintensity 8 can work as well. Lower number, less stale shares, but can hurt performance on some rigs.
|
|
|
We still haven't bought a LHR2 card. The LHR1 seem to unlock, but not always. Buying amd cards seems to be the best option right now... Enable Smart Access Memory ...
|
|
|
What will the default xintensity value be for Nvidia cards? On another note. I am currently running xintensity 1025/dagintensity 9 for 3060 TI’s FE for more than 12+ hours with 0 reject shares.
Same intensity but different, you will see tomorrow. The xintensity is pool dependent. Flexpool and crazpool will not work with xintensity 1025. (4% stale shares)
|
|
|
1.23 will have a tiny speedup in the cuda 11.5 build for nvidia. For AMD less stale shares on etheremine, flexpool, crazypool and nicehash. We use --xintensity 24 as default on these pools instead of -1.
v1.23 will be released tomorrow
|
|
|
My latest and most stable TBM 1.22 run -- photos and a short video to show that I'm not photoshopping or showing very unstable results as being useful. https://imgur.com/a/stWmdM3Nice screenshots, thanks for testing. You should post the video on youtube.
|
|
|
sp_, i found probs yesterday.. --lock-cclock [[1300,1300],[1390,1390],[1390,1390],[0,0]]
Fixed in v1.23
|
|
|
Yes, you are right. One of them has to be wifi in a remote place for cable. I tried reset, unfortunately it didn't work
ssh into the rig and run: $ sudo apt install libcurl4-openssl-dev
|
|
|
How can we find the best --xintensity and --dagintensity? I have some 6800xt, 3070, 3080. Which should be the best choices? Based on what?
- High --xintensity, higher amout of stale shares, higher speed. Some pools doesn't pay for stales so the setting is pool dependent. (AMD intensity settings are 1/4 of NVIDIA) - --dagintensity , 1-8(9) is for high memclocks. When dag buffer crash. 1 is low (slow dag) 8 amd is the fastest, 9 on nvidia. or 0 (default) - --kernel , Different kernels can give a speedup on different models. Your power setting can also count. Autotune will try to find the fastest kernel for you.
|
|
|
Make sure that your wifi router doesn't have a local proxy server. You need to correct directly to the pool. Also don't use ssl connection to the pool because the miner has issues with it. Use normal the ports.
|
|
|
I try mining with unMineable and encountered this problem - If I use the Terminal to check the ping to the server ethash-eu.unminable.com avg ping 54 and in the miner the ping is very high 399 It's the same with nanopool What might be the problem? Not sure we can check it. Can you open a ticket on our github? https://github.com/sp-hash/TeamBlackMiner/issuesMake sure that you are connecting to the eu server in the launch config. ethash.unmineable.com --port 3333 ethash-us.unmineable.com --port 3333 ethash-ca.unmineable.com --port 3333 ethash-eu.unmineable.com --port 3333 ethash-asia.unmineable.com --port 3333
|
|
|
My last question is I get high mhs on first run. After a few minutes the mhs goes down and stays constant, any solution?
I see that the coreclock of your gpus are all different. It's probobly because they trottle. Not enough power or too much heat. You can increase the templimit, but not sure if this is the problem.
|
|
|
|