Team Black Miner (ETHW ETC Vertcoin Ravencoin Zilliqa +dual +tripple mining ) - page 88.

BitBlitzer

newbie

Activity: 24

Merit: 0

Yea I have already broken that!

I have a few cards with aluminium shims instead of thermal pads and can max the memory to +2000. Gives around 111.5MH/s (reported speed). Memory temps drop like crazy with aluminium/copper shims.

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: BitBlitzer on October 18, 2021, 03:00:01 PM

So all things being equal, is a kernel 7 (TBM 1.16) on a RTX 3080 faster compared to whatever kernel was used in 1.15?

no. The cuda 11.2 build has been optimized.
In v1.16 cuda 11.4 #6 is faster. Or it depends on your launch config. There was a user who asked for 110MHASH. Why don't you try

If you run at 50% tdp, the program will select the fastest kernel for your config automaticly.

BitBlitzer

newbie

Activity: 24

Merit: 0

So all things being equal, is a kernel 7 (TBM 1.16) on a RTX 3080 faster compared to whatever kernel was used in 1.15?

Is it faster in reported MH/s or in actual shares?

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

The reported hashrate on ethermine is working fine here. Why don't you share your realtime mining link so we can investigate.

gtx 1660 with memtweak (bios mod) ~35MHASH !

--kernel 0 / GPUMemoryTransferRateOffsetAllPerformanceLevels=1700

57m [2021-10-18 17:57:42.023] GPU0 1660S 0.00 H/W 34.98 MH/s 35.01 MH/s 18/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU1 1660S 0.00 H/W 35.14 MH/s 35.11 MH/s 29/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU2 1660S 0.00 H/W 35.18 MH/s 35.10 MH/s 22/0/1 (100.00%)
57m [2021-10-18 17:57:42.023] GPU3 1660S 0.00 H/W 35.10 MH/s 35.01 MH/s 22/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU4 1660S 0.00 H/W 33.33 MH/s 33.49 MH/s 24/0/1 (100.00%)
57m [2021-10-18 17:57:42.023] GPU5 1660S 0.00 H/W 33.43 MH/s 33.47 MH/s 21/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU6 1660S 0.00 H/W 35.02 MH/s 34.96 MH/s 26/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU7 1660S 0.00 H/W 34.83 MH/s 34.96 MH/s 32/0/1 (100.00%)
57m [2021-10-18 17:57:42.023] GPU8 1660S 0.00 H/W 35.03 MH/s 35.03 MH/s 20/0/1 (100.00%)
57m [2021-10-18 17:57:42.023] GPU9 1660S 0.00 H/W 35.03 MH/s 35.03 MH/s 24/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU10 1660S 0.00 H/W 35.09 MH/s 35.01 MH/s 32/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU11 1660S 0.00 H/W 34.94 MH/s 35.00 MH/s 29/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU12 1660S 0.00 H/W 32.47 MH/s 32.42 MH/s 25/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] GPU13 1660S 0.00 H/W 35.16 MH/s 35.09 MH/s 30/0/0 (100.00%)
57m [2021-10-18 17:57:42.023] 0.00 H/W 484.71 MH/s 484.68 MH/s 354/0/4 (100.00%)

--kernel 0 low power,cards with few shaders, low intensity
..
--kernel 7 higher power good on cards with many shaders, high intensity

If your rig has low tdp, low intensity, autotune will probobly select a lower number kernel.
The rtx 3060 and up works best on 5-7
The gtx 1070 has a new record speed on kernel #4

[moderator's note: consecutive posts merged]

matmator

sr. member

Activity: 487

Merit: 252

bou !

Hello,

why we always reported hashrate = 0 on ethermine?

I think that with the maj ca going to be arranged but no!

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: sp_ on October 17, 2021, 02:42:15 PM

Why don't you do a comparison between tbminer at 108.5 MHASH and tbminer at 100 MHASH. To reach 100MHASH you need to nerfe the --xintensity, cut alot of power, reduce the clocks and run on a pool that doesn't pay for stale work. The cuda 11_2 build is also helping to reduce the hashrate.

Ramsete2

jr. member

Activity: 87

Merit: 5

Quote from: sp_ on October 18, 2021, 10:47:27 AM

Quote from: Ramsete2 on October 18, 2021, 10:03:19 AM

At the moment I had to use --xintensity 48 for AMD cards.
Setting --xintensity -1 will cause too many stales for me.

Why do you mine on a pool that doesn't pay for stales?

because it's still the most profitable for me.

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: Ramsete2 on October 18, 2021, 10:03:19 AM

At the moment I had to use --xintensity 48 for AMD cards.
Setting --xintensity -1 will cause too many stales for me.

Why do you mine on a pool that doesn't pay for stales?

Ramsete2

jr. member

Activity: 87

Merit: 5

Quote from: ?? on ??

Quote from: Ramsete2 on October 17, 2021, 12:48:48 PM

Problem with RIG full of AMD RX6800, no statistics and wrong name of cards.
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU0 gfx1030 0.00 H/W 65.04 MH/s 63.93 MH/s 1/0/0 (100.00%)

v1.16 should be faster than v1.15 on AMD with --xintensity -1 . We need to rewrite the way we get AMD stats, so this in the todo list.

At the moment I had to use --xintensity 48 for AMD cards.
Setting --xintensity -1 will cause too many stales for me.

GMIner VS TBMiner, update after 20 hours
It seems that GMiner overtook the race...

TBMiner (ver 1.15)
ID BOARD HASHRATE/W HASHRATE AVERAGE SHARES
GPU0 3080 0.00 H/W 102.39 MH/s 102.40 MH/s 797/0/7 (100.00%)
GPU1 3080 0.00 H/W 102.70 MH/s 102.70 MH/s 809/0/7 (100.00%)
GPU2 3080 0.00 H/W 102.69 MH/s 102.69 MH/s 843/0/10 (100.00%)
GPU3 3080 0.00 H/W 102.68 MH/s 102.69 MH/s 818/0/6 (100.00%)
GPU4 3080 0.00 H/W 102.63 MH/s 102.69 MH/s 753/0/3 (100.00%)
GPU5 3080 0.00 H/W 102.72 MH/s 102.69 MH/s 845/0/7 (100.00%)
0.00 H/W 615.83 MH/s 615.87 MH/s 4865/0/40 (100.00%)

GMiner (ver 2.70)
+---+-----+------------+---------+--------+---------------+
| ID GPU Speed Shares Power Efficiency |
+---+-----+------------+---------+--------+---------------+
| 0 3080 98.42 MH/s 833/0/0 223 W 441.34 KH/W |
| 1 3080 98.83 MH/s 792/0/0 217 W 455.43 KH/W |
| 2 3080 97.78 MH/s 806/0/0 222 W 440.45 KH/W |
| 3 3080 98.83 MH/s 811/0/0 221 W 447.18 KH/W |
| 4 3080 98.82 MH/s 810/0/0 218 W 453.28 KH/W |
| 5 3080 98.83 MH/s 820/0/0 222 W 445.17 KH/W |
| 591.50 MH/s 4872/0/0 1323 W 447.09 KH/W |
+---+-----+------------+---------+--------+---------------+

[moderator's note: consecutive posts merged]

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

v1.16 should be the fastest ethereum miner on most cards.

Quote from: Jeggu on October 15, 2021, 12:51:25 PM

Any plans on optimizing the miner for LHR cards? I've got a couple 3080 tis looking for a home.

Yes. Which of teams has the fastest LHR kernel ?

Quote from: Ramsete2 on October 17, 2021, 12:48:48 PM

Problem with RIG full of AMD RX6800, no statistics and wrong name of cards.
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU0 gfx1030 0.00 H/W 65.04 MH/s 63.93 MH/s 1/0/0 (100.00%)

v1.16 should be faster than v1.15 on AMD with --xintensity -1 . We need to rewrite the way we get AMD stats, so this in the todo list.

Quote from: hashatharias on October 16, 2021, 09:36:22 PM

The core lock is really a blessing, locking the core with afterburner is such a chore... What would also be quite nice is the way to set the memory frequency and fan speed as well.

--lock-mclock has been added in v1.16.

[moderator's note: consecutive posts merged]

KillrBee

jr. member

Activity: 139

Merit: 3

Quote from: sp_ on October 17, 2021, 12:07:59 PM

Quote from: KillrBee on October 17, 2021, 08:44:52 AM

There seems to be a bug in the "Average" hash rate when starting up AMD RNDA2 cards and a high intensity value.
https://imgur.com/a/382cFXO

yes, this bug is also causing the xintensity -1 to fail on some rigs. Fixed in v1.16. What's up with the colors? Agent Orange Huh

?

It's Radeon Red Oktober! / oктябpe

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: BitBlitzer on October 17, 2021, 12:39:01 PM

Out of the total shares TBM has 125 stale shares so that is about 2%

In 1.16 we will reduce the stales to around 1% or less on etheremine and flexpool.

v1.16
1. Added new and faster nvida CUDA kernels. Improved performance with auto-tune to find the best option for each card.
2. Changed commandline long options use of underscore. Replaced with dash. See --help.
3. Added --kernel option. (NVIDIA) 8 different kernels to choose from. Without this set, autotune is enabled.
4. Added --lock-mclock option. Specifies minimum and maximum memory clocks per GPU. (run as admin)
5. Added --power-limit option. Specifies maximum power limit in watts per GPU. (run as admin)
6. --xintensity reduced to 192 to remove stales from pools.
7. Fixed a bug in average hashrate and xintensity -1 on amd cards.

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

TeamBlackMiner_1_16_cuda_11_2.7z
https://www.virustotal.com/gui/file/ca5cece84e34d09ad4748c8eb760e2577823d128cea1558f90090d1e2b26d6b1?nocache=1

TeamBlackMiner_1_16_cuda_11_4.7z
https://www.virustotal.com/gui/file/5614fe1e79b9a440ab3892cfc5036b1ff3e6c44e4f2cda146298760aebbe4e88?nocache=1

TeamBlackMiner_1_16_Ubuntu_18_04_Cuda_11_4
https://www.virustotal.com/gui/file/4a79adfdd1ae321e9fa9e66cf84dcdcbb1b9fff137b897822157db6eecc3b7ff?nocache=1

[moderator's note: consecutive posts merged]

Ramsete2

jr. member

Activity: 87

Merit: 5

Quote

Usually this is a driver problem. Try to reinstall the driver.
What happens when you type $ nvidia-smi in the shell ?

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: Ramsete2 on October 18, 2021, 02:37:02 AM

Preliminary comparison after 12 hours: GMiner VS TBMiner
Used two gemini rigs of 6x3080, identical overclocks (Xintensity 192). Crazypool
591.50 MH/s 2682/0/0
TBMiner (ver 1.15):
615.98 MH/s 2804/0/27

Crazypool doesn't pay for stales anymore. Interesting find with --xintensity 192..

It's Easy to compare by counting the number of accepted shares.

2804/2682 = +4.54% poolside profit. v1.16 will be faster.

Quote from: Ramsete2 on October 18, 2021, 02:37:02 AM

SP, any news about the problem of zero statistics shown?

Usually this is a driver problem. Try to reinstall the driver.
What happens when you type $ nvidia-smi in the shell ?

Quote from: Ramsete2 on October 18, 2021, 02:37:02 AM

Xintensity 192

Tested a rig with mixed cards v1.16 with autotune on --xintensity 192. (181.80MHASH). Same speed as 1.15 with --xintensity 400.
v1.16 with autotune and xintensity 400 does 184mhash.

Ramsete2

jr. member

Activity: 87

Merit: 5

Preliminary comparison after 12 hours: GMiner VS TBMiner
Used two gemini rigs of 6x3080, identical overclocks (Xintensity 192). Crazypool

GMiner (ver 2.70):
+---+-----+------------+---------+--------+--------------+
| ID GPU Speed Shares Power Efficiency |
+---+-----+------------+---------+--------+--------------+
| 0 3080 98.42 MH/s 449/0/0 223 W 441.35 KH/W |
| 1 3080 98.83 MH/s 466/0/0 215 W 459.67 KH/W |
| 2 3080 97.78 MH/s 453/0/0 220 W 444.46 KH/W |
| 3 3080 98.83 MH/s 436/0/0 220 W 449.22 KH/W |
| 4 3080 98.82 MH/s 455/0/0 216 W 457.48 KH/W |
| 5 3080 98.83 MH/s 423/0/0 220 W 449.22 KH/W |
| 591.50 MH/s 2682/0/0 1314 W 450.15 KH/W |
+---+-----+------------+---------+--------+--------------+

TBMiner (ver 1.15):
ID BOARD HASHRATE/W HASHRATE AVERAGE SHARES
GPU0 3080 0.00 H/W 102.42 MH/s 102.40 MH/s 447/0/4 (100.00%)
GPU1 3080 0.00 H/W 102.68 MH/s 102.72 MH/s 464/0/6 (100.00%)
GPU2 3080 0.00 H/W 102.78 MH/s 102.74 MH/s 502/0/6 (100.00%)
GPU3 3080 0.00 H/W 102.75 MH/s 102.74 MH/s 486/0/4 (100.00%)
GPU4 3080 0.00 H/W 102.67 MH/s 102.71 MH/s 418/0/1 (100.00%)
GPU5 3080 0.00 H/W 102.68 MH/s 102.71 MH/s 487/0/6 (100.00%)
0.00 H/W 615.98 MH/s 616.02 MH/s 2804/0/27 (100.00%)

I'll try to reduce a bit more Xintensity to eliminate stales.

SP, any news about the problem of zero statistics shown?

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Here is a preview of the autotune function in Team Black Miner v1.16. Tests have shown around +5% on low power configurations and overall a few percent faster.
NVIDIA only for now...

Will perhaps release tonight or tomorrow.

ayiphelmy

copper member

Activity: 416

Merit: 105

Quote from: ?? on ??

Quote from: ayiphelmy on October 17, 2021, 10:26:45 AM

let see if this version can beat nbminer on 1070ti, v1.15 still 1.5 mhs slower

We can Memtweak too. v1.16 will have 8 cuda kernels, you need to test them all to find the fastest one.

--kernel parameter

Any more ethereum miners we need to checkout? What's the fastest on the 1070ti?

Already tested claymore, t-rex, gminer, nbminer, lolminer, phoenixminer, srbminer, teamredminer

Claymore seems to be the most popular one..

34.24 mhs 1070ti with nbminer

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: ZerodayTrader on October 17, 2021, 01:28:56 PM

Why don't you do a comparison between tbminer at 108.5 MHASH and tbminer at 100 MHASH. To reach 100MHASH you need to nerfe the --xintensity, cut alot of power, reduce the clocks and run on a pool that doesn't pay for stale work. The cuda 11_2 build is also helping to reduce the hashrate.

ZerodayTrader

newbie

Activity: 29

Merit: 0

Quote from: BitBlitzer on October 17, 2021, 12:39:01 PM

Out of the total shares TBM has 125 stale shares so that is about 2% (125/5700)

Try out with High Difficulty Port. Otherwise; switch the Pool, or reduce Memory OC. On Binance ( Roll Eyes

) i had >2% staleshares too, and now mining somewhere else Shiba INU for ETHash, and guess what happens,..

https://i.imgur.com/GcSYCOH.png

Also same effect for my 1070.

https://i.imgur.com/LUpa0AH.png

Ramsete2

jr. member

Activity: 87

Merit: 5

Problem with RIG full of AMD RX6800, no statistics and wrong name of cards.
The statistics bug is present also on NVidia RTX rigs, can you solve it?

│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] crazypool.org (ethash) PING: 20ms DIFFICULTY: 2.00 EPOCH: 447
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028]
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028]
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] ID BOARD TYPE TEMP FAN CORE MEM WATT kW/h COST/h
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU0 gfx1030 OpenCL 0/0 0 0 0 0 0.00 0.00USD
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU1 gfx1030 OpenCL 0/0 0 0 0 0 0.00 0.00USD
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU2 gfx1030 OpenCL 0/0 0 0 0 0 0.00 0.00USD
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU3 gfx1030 OpenCL 0/0 0 0 0 0 0.00 0.00USD
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU4 gfx1030 OpenCL 0/0 0 0 0 0 0.00 0.00USD
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU5 gfx1030 OpenCL 0/0 0 0 0 0 0.00 0.00USD
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU6 gfx1030 OpenCL 0/0 0 0 0 0 0.00 0.00USD
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] 0 0.00 0.00USD
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028]
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] ID BOARD HASHRATE/W HASHRATE AVERAGE SHARES
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU0 gfx1030 0.00 H/W 65.04 MH/s 63.93 MH/s 1/0/0 (100.00%)
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU1 gfx1030 0.00 H/W 65.45 MH/s 64.08 MH/s 0/0/0 (100.00%)
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU2 gfx1030 0.00 H/W 65.70 MH/s 63.97 MH/s 3/0/0 (100.00%)
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU3 gfx1030 0.00 H/W 65.45 MH/s 63.95 MH/s 2/0/0 (100.00%)
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU4 gfx1030 0.00 H/W 65.44 MH/s 63.83 MH/s 0/0/0 (100.00%)
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU5 gfx1030 0.00 H/W 69.01 MH/s 63.78 MH/s 2/0/0 (100.00%)
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] GPU6 gfx1030 0.00 H/W 65.00 MH/s 63.34 MH/s 2/0/1 (100.00%)
│[2021-10-17 17:44:47][info][miner/0]: 3m [2021-10-17 17:44:46.028] 0.00 H/W 461.10 MH/s 446.89 MH/s 10/0/1 (100.00%)

Topic: Team Black Miner (ETHW ETC Vertcoin Ravencoin Zilliqa +dual +tripple mining ) - page 88. (Read 35053 times)