⛏Team Black Miner (ETHB3 ETHW ETC VTC KAWPOW FIROPOW EVRPROGPOW MEOWPOW +dual)

[sp_]

1.45 has a gpu timeout bug on AMD cards on high intensity. Upgrade to 1.47

[Salva52]

What do you recommand for a 3080FE and 3090Fe on miningpoolhub ? (xintensity/tweak/kernel) Thanks

[sp_]

What do you recommand for a 3080FE and 3090Fe on miningpoolhub ? (xintensity/tweak/kernel) Thanks

rtx 3090 --xintensity 3000 --kernel 1 or 12. Run the autotune a few times to verify. The best kernel depends on your power settings.
rtx 3080 --xintensity 2500? --kernel 1 or 12. Run the autotune a few times to verify. The best kernel depends on your power settings.

 try without tweak first, if you get stale shares incease the tweak to 4 or more

[atoue]

1.45 has a gpu timeout bug on AMD cards on high intensity. Upgrade to 1.47

Hi, I find that in 1.47 GPU timed out error still exist

https://i.ibb.co/pjWXGHC/Capture-5-LI.jpg

[Salva52]

What do you recommand for a 3080FE and 3090Fe on miningpoolhub ? (xintensity/tweak/kernel) Thanks

rtx 3090 --xintensity 3000 --kernel 1 or 12. Run the autotune a few times to verify. The best kernel depends on your power settings.
rtx 3080 --xintensity 2500? --kernel 1 or 12. Run the autotune a few times to verify. The best kernel depends on your power settings.

 try without tweak first, if you get stale shares incease the tweak to 4 or more

Thanks sp_
I will try this setup.
But stale shares are bad ? Why ? If they accepted by the pool (mph)

[sp_]

Thanks sp_
I will try this setup.
But stale shares are bad ? Why ? If they accepted by the pool (mph)

if you see in the  Round Shares Submitted shares , you sometimes see stale shares. Level1 stales are payed fully, but lower lewels give a lower payout I think.

On 2miners stale shares can give rejected shares, but you need a really high xintensity to trigger that. 8000 or more.

[Salva52]

Ah ok, sorry, i haven't understood.
Yes i have "1 level stale 12.14%" for.now, i will watch if level change. But that's sure 1 level stales are fully payed ?
Thanks

[atoue]

I got a good result with TBM for more than 24 hours, but the GPU timed out in AMD GPU came back and makes the result poor again

[sp_]

I got a good result with TBM for more than 24 hours, but the GPU timed out in AMD GPU came back and makes the result poor again

From the picture it looks like your gpu has crashed. Both temps and power indicate that. Run with --no-ansi. and lower the memclock -10. High intensity might need a slightly lower clock to run stable.

In the bug fixed, both power and temps where high when it timed out, which was not a real crash, but a glitch in the gpu test code.

[atoue]

I got a good result with TBM for more than 24 hours, but the GPU timed out in AMD GPU came back and makes the result poor again

From the picture it looks like your gpu has crashed. Both temps and power indicate that. Run with --no-ansi. and lower the memclock -10. High intensity might need a slightly lower clock to run stable.

In the bug fixed, both power and temps where high when it timed out, which was not a real crash, but a glitch in the gpu test code.

Thanks!

What does --xintensity [-1] do?

[sp_]

Thanks!

What does --xintensity [-1] do?

It will dynamically allocate the workload to the gpu based on the hashrate. So it's a floating intensity and not a fixed one.

[KillrBee]

I'm still getting a number of timeouts and random crashes on 1.47 with the AMD and Nvidia cards running together in a single miner instance.  Would there be any reason why having the two in a single instance would cause instability or reduced performance?  HiveOS doesn't keep historical crash reports, but here are the last couple of lines from the ones I was able to capture:


send buffer: '(null)'
send buffer length: 146
recv buffer: '(null)'
recv buffer length: 213
recv buffer more data flag: 2058230960
-------------------------

recv buffer length: 211
recv buffer more data flag: -1564439376
My instances are all running standard OC/UV and are actually running at a much lower memory clock than I was previously using:


00:00:25 [2022-01-30 22:13:44.974] us-eth.2miners.com (ethash)    PING: 46ms    VARDIFF: 2.03     EPOCH: 470
00:00:25 [2022-01-30 22:13:44.974]                                                                                
00:00:25 [2022-01-30 22:13:44.974]
00:00:25 [2022-01-30 22:13:44.974] ID      BOARD    TYPE    KERN    XINT    TEMP    FAN     CORE    MEM     WATT  
00:00:25 [2022-01-30 22:13:44.974] GPU0    gfx1030  OpenCL  0       1000    30/0    73      1350    1059    122    
00:00:25 [2022-01-30 22:13:44.974] GPU1    gfx1030  OpenCL  0       1000    25/0    73      1325    1059    112    
00:00:25 [2022-01-30 22:13:44.974] GPU2    gfx1030  OpenCL  0       1000    25/0    73      1365    1059    134    
00:00:25 [2022-01-30 22:13:44.974] GPU3    gfx1030  OpenCL  0       1000    26/0    73      1350    1059    111    
00:00:25 [2022-01-30 22:13:44.974] GPU4    3090      CUDA    12       4000    33/0    85      1590    11425  331    
00:00:25 [2022-01-30 22:13:44.974] GPU5    2060S    CUDA    12       4000    28/0    85      1455    7975    102    
00:00:25 [2022-01-30 22:13:44.974]                                                                                                       912    


And for some reason, there is always one of my AMD cards (not consistently the same one) that reports a good hashrate, but the share rate calculation results in around 10 MH/s lower than reported:


00:06:11 [2022-01-31 09:21:28.082] us-eth.2miners.com (ethash)    PING: 82ms    VARDIFF: 2.03     EPOCH: 470
00:06:11 [2022-01-31 09:21:28.082]                                                                                
00:06:11 [2022-01-31 09:21:28.082]
00:06:11 [2022-01-31 09:21:28.082] ID      BOARD    TYPE    KERN    XINT    TEMP    FAN     CORE    MEM     WATT  
00:06:11 [2022-01-31 09:21:28.082] GPU0    gfx1030  OpenCL  0       1000    29/0    73      1350    1059    124    
00:06:11 [2022-01-31 09:21:28.082] GPU1    gfx1030  OpenCL  0       1000    25/0    73      1325    1059    114    
00:06:11 [2022-01-31 09:21:28.082] GPU2    gfx1030  OpenCL  0       1000    24/0    73      1365    1059    135    
00:06:11 [2022-01-31 09:21:28.082] GPU3    gfx1030  OpenCL  0       1000    25/0    73      1350    1059    114    
00:06:11 [2022-01-31 09:21:28.082] GPU4    3090       CUDA    2       4000    33/0    0        0          0         337    
00:06:11 [2022-01-31 09:21:28.082] GPU5    2060S    CUDA    13      4000    28/0    85      1455    7975    103    
00:06:11 [2022-01-31 09:21:28.082]                                                                          927    
00:06:11 [2022-01-31 09:21:28.082]
00:06:11 [2022-01-31 09:21:28.082] ID      BOARD    HASHRATE/W    HASHRATE      AVERAGE       SHARES       RATE    
00:06:11 [2022-01-31 09:21:28.082] GPU0    gfx1030  529.70 kH/W   65.68 MH/s    65.64 MH/s    151/2/107    98.69%  
00:06:11 [2022-01-31 09:21:28.082] GPU1    gfx1030  581.25 kH/W   65.68 MH/s    65.65 MH/s    175/1/112    99.43%  
00:06:11 [2022-01-31 09:21:28.082] GPU2    gfx1030  538.80 kH/W   65.20 MH/s    65.24 MH/s    132/0/76     100.00%
00:06:11 [2022-01-31 09:21:28.082] GPU3    gfx1030  581.35 kH/W   65.69 MH/s    65.64 MH/s    152/1/91     99.35%  
00:06:11 [2022-01-31 09:21:28.082] GPU4    3090     399.54 kH/W   134.64 MH/s   134.65 MH/s   356/0/61     100.00%
00:06:11 [2022-01-31 09:21:28.082] GPU5    2060S    425.85 kH/W   44.29 MH/s    44.24 MH/s    126/3/25     97.67%  
00:06:11 [2022-01-31 09:21:28.082]                  509.42 kH/W   441.18 MH/s   441.06 MH/s   1092/7/472   99.36%  
00:06:11 [2022-01-31 09:21:28.082]                                                                                
00:06:11 [2022-01-31 09:21:28.082] SHARES PER MINUTE: 2.95            POOL HASHRATE: 429.23 MH/s


Speed Calculation GPU 2:
132 shares / 371 seconds = 0.356 shares/second ::  0.356 * 145 MH/s = 51.5 MH/s


Another longer run:


00:17:27 [2022-01-25 08:17:42.682]
00:17:27 [2022-01-25 08:17:42.682] us-eth.2miners.com (ethash)    PING: 25ms    VARDIFF: 2.03     EPOCH: 469
00:17:27 [2022-01-25 08:17:42.682]                                                                                 
00:17:27 [2022-01-25 08:17:42.682]
00:17:27 [2022-01-25 08:17:42.682] ID      BOARD    TYPE    KERN    XINT    TEMP    FAN     CORE    MEM     WATT   
00:17:27 [2022-01-25 08:17:42.682] GPU0    gfx1030  OpenCL  0       655     24/0    73      1350    1066    125   
00:17:27 [2022-01-25 08:17:42.682] GPU1    gfx1030  OpenCL  0       655     19/0    73      1325    1066    114   
00:17:27 [2022-01-25 08:17:42.682] GPU2    gfx1030  OpenCL  0       655     19/0    73      1365    1066    137   
00:17:27 [2022-01-25 08:17:42.682] GPU3    gfx1030  OpenCL  0       655     20/0    73      1350    1066    114   
00:17:27 [2022-01-25 08:17:42.682] GPU4    3090     CUDA    8       5801    26/0    0       0       0       336   
00:17:27 [2022-01-25 08:17:42.682] GPU5    2060S    CUDA    12      3000    27/0    85      1650    7950    115   
00:17:27 [2022-01-25 08:17:42.682]                                                                          941   
00:17:27 [2022-01-25 08:17:42.682]
00:17:27 [2022-01-25 08:17:42.682] ID      BOARD    HASHRATE/W    HASHRATE      AVERAGE       SHARES       RATE   
00:17:27 [2022-01-25 08:17:42.682] GPU0    gfx1030  528.39 kH/W   66.05 MH/s    65.96 MH/s    388/0/173    100.00%
00:17:27 [2022-01-25 08:17:42.682] GPU1    gfx1030  574.01 kH/W   66.01 MH/s    65.99 MH/s    429/1/199    99.77% 
00:17:27 [2022-01-25 08:17:42.682] GPU2    gfx1030  472.20 kH/W   65.64 MH/s    65.55 MH/s    456/0/175    100.00%
00:17:27 [2022-01-25 08:17:42.682] GPU3    gfx1030  584.50 kH/W   66.05 MH/s    65.96 MH/s    448/0/210    100.00%
00:17:27 [2022-01-25 08:17:42.682] GPU4    3090     397.74 kH/W   133.64 MH/s   133.63 MH/s   928/0/251    100.00%
00:17:27 [2022-01-25 08:17:42.682] GPU5    2060S    388.06 kH/W   45.01 MH/s    45.03 MH/s    334/6/69     98.24% 
00:17:27 [2022-01-25 08:17:42.682]                  490.81 kH/W   442.40 MH/s   442.12 MH/s   2983/7/1077  99.77% 
00:17:27 [2022-01-25 08:17:42.682]                                                                                 
00:17:27 [2022-01-25 08:17:42.682] SHARES PER MINUTE: 2.85            POOL HASHRATE: 414.36 MH/s


388 / 1047 = 0.37 * 145 = 53.7 MH/s
There's always 1 AMD card that seems to be running slow at the pool, but fine in the client.

[sp_]

00:17:27 [2022-01-25 08:17:42.682] GPU4    3090     CUDA    8       5801    26/0    0       0       0       336    
00:17:27 [2022-01-25 08:17:42.682] GPU5    2060S    CUDA    12      3000    27/0    85      1650    7950    115    

Are you sure this is the 1.47 version? A bug was fixed in v1.45 that fix the memclock/coreclock on mixed rigs. but your rtx 3090 is 0.
v1.48 will have some more output to console if the gpu is lagging. +less cpu usage, and LHR unlock support for the RTX 3050.

The Nvidia solver is running with blocking sync, that might stop the cl code to perform optimal.

[KillrBee]

00:17:27 [2022-01-25 08:17:42.682] GPU4    3090     CUDA    8       5801    26/0    0       0       0       336    
00:17:27 [2022-01-25 08:17:42.682] GPU5    2060S    CUDA    12      3000    27/0    85      1650    7950    115    

Are you sure this is the 1.47 version? A bug was fixed in v1.45 that fix the memclock/coreclock on mixed rigs. but your rtx 3090 is 0

Yes it's v. 1.47 and the 0's are because Nvidia cards that are highly overclocked will glitch on the sensors in HiveOS.

I'm going to reduce the intensities and see if that can increase the stability.  I'll let you know if there are any changes in behaviour.

[sp_]

v1.48
1. LHR partial unlock on the RTX 3050.
2. Improved the ETHProxy implementation. Fewer rejected shares.
3. Added LHR reset counter to output.
4. Less cpu usage and more stable LHR.
5. Add gpu timout warning in the output for testing purposes.


https://github.com/sp-hash/TeamBlackMiner/releases

TeamBlackMiner_1_48_cuda_11_4.7z
https://www.virustotal.com/gui/file/5f1ed688d37f20510ad980d91addb449101e5a920c41ef241fb417276853e684?nocache=1

TeamBlackMiner_1_48_cuda_11_5.7z
https://www.virustotal.com/gui/file/689662e8f17f01a04f5f88f65bb724033059750eaa9d180ea05f4070e6b578f7?nocache=1

TeamBlackMiner_1_48_Ubuntu_18_04_Cuda_11_4.tar.xz
https://www.virustotal.com/gui/file/5fe33315dc81f51b854f730309f728ae286cd54aec3bc7d52e9eed10a4047eba?nocache=1

TeamBlackMiner_1_48_Ubuntu_18_04_Cuda_11_5.tar.xz
https://www.virustotal.com/gui/file/4e7e11bc9097a1b2634184eedf9b64522d7a68d174e3206571234d3c660a5973?nocache=1

[atoue]

00:17:27 [2022-01-25 08:17:42.682] GPU4    3090     CUDA    8       5801    26/0    0       0       0       336    
00:17:27 [2022-01-25 08:17:42.682] GPU5    2060S    CUDA    12      3000    27/0    85      1650    7950    115    

Are you sure this is the 1.47 version? A bug was fixed in v1.45 that fix the memclock/coreclock on mixed rigs. but your rtx 3090 is 0

Yes it's v. 1.47 and the 0's are because Nvidia cards that are highly overclocked will glitch on the sensors in HiveOS.

I'm going to reduce the intensities and see if that can increase the stability.  I'll let you know if there are any changes in behaviour.

any updates? I have similar issue and reducing the intensities and also oc doesn't work on me, the time out event still coming back lol. With other miner it runs rock stable

[sp_]

any updates? I have similar issue and reducing the intensities and also oc doesn't work on me, the time out event still coming back lol. With other miner it runs rock stable

Try the new version 1.48. My mixed card testrig with 7 cards. AMD / NVIDIA ran all night without any timeouts. and around +16MHASH faster poolside hash.

[KillrBee]

any updates? I have similar issue and reducing the intensities and also oc doesn't work on me, the time out event still coming back lol. With other miner it runs rock stable

Try the new version 1.48. My mixed card testrig with 7 cards. AMD / NVIDIA ran all night without any timeouts. and around +16MHASH faster poolside hash.

I reduced the intensities (from 1000 to 800) on my AMD cards and it's run for just over 11 hours on HiveOS (using 1.47) and 10 hours on windows using 1.48.  Before I lowered my intensity from 1000 to 750 on the windows machine, I was getting a lot of "the GPU is responding slowly" and it would stop accepting jobs for about 5-6 minutes before finally timing out the card.

With HiveOS, I would just get the drop in power of the card, but the hash-rate would remain the same, it would also stop accepting jobs eventually (although the Nvidia cards were still running properly, and the client would continue to send in accepted shares for 3-4 minutes after the client stopped accepting jobs; which makes me curious about what the time threshold is for providing stale shares; none of these were rejected by the pool.

Overall hashrate has reduced a bit at the pool, especially on the Windows client where I was running at 66-67 MH/s on 15+ hour runs, while I'm sitting around 62 MH/s after 10 hours right now.  (difference was intensity of 1000 vs. 750).  

[sp_]

I was getting a lot of "the GPU is responding slowly" and it would stop accepting jobs for about 5-6 minutes before finally timing out the card.

I managed to reproduce this now on my mixed card rig with the AMD cards running on miningpoolhub with --xintensity 1024. Added logging and debug, and let it run over night.

[atoue]

I was getting a lot of "the GPU is responding slowly" and it would stop accepting jobs for about 5-6 minutes before finally timing out the card.

I managed to reproduce this now on my mixed card rig with the AMD cards running on miningpoolhub with --xintensity 1024. Added logging and debug, and let it run over night.

can't wait to see the updates about this!