Hey guys,
I tried to disable TBM on some GPU but each time I use the option I get an error message and the whole mining software will not start.
https://ibb.co/QCmnZfz
Topic: Team Black Miner (ETHW ETC Vertcoin Ravencoin Zilliqa +dual +tripple mining ) - page 78. (Read 35053 times)
October 25, 2021, 04:55:20 AM
October 25, 2021, 04:30:09 AM
October 25, 2021, 01:08:46 AM
October 25, 2021, 12:59:10 AM
try --cl-devices [0,1,2,3,4,5,6,7] instead of --auto-detect . Don't use both options at the same time.
I just use the --cl-devices command in the "extra config arguments" on HiveOs, and it automatically don't use the --auto-detect?
October 25, 2021, 12:02:51 AM
try --cl-devices [0,1,2,3,4,5,6,7] instead of --auto-detect . Don't use both options at the same time.
October 24, 2021, 11:53:39 PM
No the 3gb is from the opencl call to determine the size of your card. Could be related to a wrong gpu platform, or a driver installation gone wrong. You should do a driver uninstall with wipe and do a reinstall.
Well, if I stop the miner and run the --list-devices command, all the devices are read again with 8gb, so it is the "free memory" what you see from the tbminer --list-devices command, not the hardware memory
October 24, 2021, 11:39:42 PM
Maybe because of the DAG generation fails? It makes sense if the memory remain "occupied" by the failed DAG generated. 8gb - DAG size, then rounded is 3gb as I see in the list.
No the 3gb is from the opencl call to determine the size of your card. Could be related to a wrong gpu platform, or a driver installation gone wrong. You should do a driver uninstall with wipe and do a reinstall.
October 24, 2021, 11:36:36 PM
Seems to be a driver problem. Weird, never seen this before..
Maybe because of the DAG generation fails? It makes sense if the memory remain "occupied" by the failed DAG generated. 8gb - DAG size, then rounded is 3gb as I see in the list.
This can also explain why I can fix only when I reboot the rig
Maybe I can open an issue on github?
October 24, 2021, 11:29:22 PM
Seems to be a driver problem. Weird, never seen this before..
October 24, 2021, 11:18:05 PM
try
tbminer --list-devices
what do you get?
tbminer --list-devices
what do you get?
Code:
Available Cuda devices
-----------------------
Available OpenCL devices
-------------------------
Device Number: 0
Device name: Ellesmere
PCIe bus id: 1
Memory: 3GB
Device Number: 1
Device name: Ellesmere
PCIe bus id: 2
Memory: 3GB
Device Number: 2
Device name: Ellesmere
PCIe bus id: 3
Memory: 3GB
Device Number: 3
Device name: Ellesmere
PCIe bus id: 4
Memory: 3GB
Device Number: 4
Device name: Ellesmere
PCIe bus id: 8
Memory: 3GB
Device Number: 5
Device name: Ellesmere
PCIe bus id: 11
Memory: 3GB
Device Number: 6
Device name: Ellesmere
PCIe bus id: 12
Memory: 3GB
Device Number: 7
Device name: Ellesmere
PCIe bus id: 13
Memory: 3GB
Device Number: 8
Device name: Ellesmere
PCIe bus id: 14
Memory: 3GB
-----------------------
Available OpenCL devices
-------------------------
Device Number: 0
Device name: Ellesmere
PCIe bus id: 1
Memory: 3GB
Device Number: 1
Device name: Ellesmere
PCIe bus id: 2
Memory: 3GB
Device Number: 2
Device name: Ellesmere
PCIe bus id: 3
Memory: 3GB
Device Number: 3
Device name: Ellesmere
PCIe bus id: 4
Memory: 3GB
Device Number: 4
Device name: Ellesmere
PCIe bus id: 8
Memory: 3GB
Device Number: 5
Device name: Ellesmere
PCIe bus id: 11
Memory: 3GB
Device Number: 6
Device name: Ellesmere
PCIe bus id: 12
Memory: 3GB
Device Number: 7
Device name: Ellesmere
PCIe bus id: 13
Memory: 3GB
Device Number: 8
Device name: Ellesmere
PCIe bus id: 14
Memory: 3GB
October 24, 2021, 11:08:57 PM
Still crashes even when running on either --kernel 6 or --kernel 4.
Does it crash with --kernel 0 ? Try to lower the clock.
On a different note, I discovered that running xintensity 768 on TBM 1.18 is stable with 0 rejects.
This is a good value for high difficulty pools that accept stales, but too high for other pools.
Some cards are recognized only 3gb. I've all 580 8gb
try
tbminer --list-devices
what do you get?
I can never get my 3090 to lock the MemClock -- nvidia-smi will say it's changed the clock, but it doesn't change anything (same with setting the fan speed through the nvidia-smi). Core clock will change, but that's it. I just add the nvidiaInspector overclocking shortcut I have into the batch file and it will run the proper memory clock from there.
We support setting the memclock too whatever the nvidia-smi tool support. This might not be as high as msi afterburner. Type this to see the supported clocks
$ nvidia-smi --query-supported-clocks=memory --format=csv
memory [MHz]
7001 MHz
6801 MHz
5001 MHz
810 MHz
405 MHz
[moderator's note: consecutive posts merged]
October 24, 2021, 11:03:16 PM
I have problem with mining eth/etc+zil over rustpool.xyz. I can´t connect to the pool.
Other miner work and etc or eth only also. Maybe TBMiner has problem with rustpool.xyz?
Other miner work and etc or eth only also. Maybe TBMiner has problem with rustpool.xyz?
rustpool.xyz seems to be down.
Weird thing with 1.19
Code:
21m [2021-10-24 21:02:47.279] POOL HASHRATE: VARDIFF
While the value was at like 2000 when I started the miner, then it switched to VARDIFFWe don't support calculating the Pool hashrate if the pool is changing difficulty while you mine. This can be implemented later, but require some more work. Which pool is this?
[moderator's note: consecutive posts merged]
October 24, 2021, 09:53:33 PM
Thanks for uploading 1.19! And all the work you've been doing 
I'm having trouble getting --lock-mclock to function correctly...--lock-cclock works perfectly (setting after dag buffer verified), but --lock-mclock just isn't working on my 3090 regardless of the input.
Have checked the formatting in the .bat etc. but just won't play nice. Afterburner works no problem and can manually set the memory clocks via nvidia-smi in an admin cmd line (albeit not able to set to anything other than the supported clocks in nvidia-smi which are rather limited in OC).
Anyone else having trouble with --lock-mclock?
I'm having trouble getting --lock-mclock to function correctly...--lock-cclock works perfectly (setting after dag buffer verified), but --lock-mclock just isn't working on my 3090 regardless of the input.
Have checked the formatting in the .bat etc. but just won't play nice. Afterburner works no problem and can manually set the memory clocks via nvidia-smi in an admin cmd line (albeit not able to set to anything other than the supported clocks in nvidia-smi which are rather limited in OC).
Anyone else having trouble with --lock-mclock?
I can never get my 3090 to lock the MemClock -- nvidia-smi will say it's changed the clock, but it doesn't change anything (same with setting the fan speed through the nvidia-smi). Core clock will change, but that's it. I just add the nvidiaInspector overclocking shortcut I have into the batch file and it will run the proper memory clock from there.
October 24, 2021, 09:12:33 PM
Even with 1.19 I've problems.
Some cards are recognized only 3gb. I've all 580 8gb
here's the screenshot
https://ibb.co/jzmTGWQ
Some cards are recognized only 3gb. I've all 580 8gb
here's the screenshot
https://ibb.co/jzmTGWQ
October 24, 2021, 08:56:36 PM
TBM v1.19 CUDA 11_4
4x 3060 TI FE
2x EVGA 3060 TI FTW ULTRA3
NVIDIA 472.12 DRIVER
4x 3060 TI FE
2x EVGA 3060 TI FTW ULTRA3
NVIDIA 472.12 DRIVER
I re-uploaded the 11_4 build. you might need to select a kernel manually with --kernel. --kernel 8 doesn't like too much oc.
on crash try
--kernel 4
Still crashes even when running on either --kernel 6 or --kernel 4.
On a different note, I discovered that running xintensity 768 on TBM 1.18 is stable with 0 rejects.
521m [2021-10-24 20:35:39.378] ID BOARD TYPE KERN XINT TEMP FAN CORE MEM WATT
521m [2021-10-24 20:35:39.378] GPU0 3060Ti Cuda 3 768 63/0 70 1416. 8077 153
521m [2021-10-24 20:35:39.378] GPU1 3060Ti Cuda 5 768 65/0 79 1676 8077 184
521m [2021-10-24 20:35:39.378] GPU2 3060Ti Cuda 2 768 68/0 70 1444 8077 153
521m [2021-10-24 20:35:39.378] GPU3 3060Ti Cuda 6 768 55/0 79 1700 8077 183
521m [2021-10-24 20:35:39.378] GPU4 3060Ti Cuda 5 768 66/0 70 1506 8077 153
521m [2021-10-24 20:35:39.378] GPU5 3060Ti Cuda 6 768 62/0 70 1446 8077 153
521m [2021-10-24 20:35:39.378] 979
521m [2021-10-24 20:35:39.378]
521m [2021-10-24 20:35:39.378] ID BOARD HASHRATE/W HASHRATE AVERAGE SHARES
521m [2021-10-24 20:35:39.378] GPU0 3060Ti 424.63 kH/W 64.97 MH/s 64.96 MH/s 236/0/6 (100.00)
521m [2021-10-24 20:35:39.378] GPU1 3060Ti 353.08 kH/W 64.97 MH/s 64.96 MH/s 226/0/6 (100.00)
521m [2021-10-24 20:35:39.378] GPU2 3060Ti 424.62 kH/W 64.97 MH/s 64.96 MH/s 237/0/2 (100.00)
521m [2021-10-24 20:35:39.378] GPU3 3060Ti 355.00 kH/W 64.97 MH/s 64.96 MH/s 241/0/3 (100.00)
521m [2021-10-24 20:35:39.378] GPU4 3060Ti 424.57 kH/W 64.96 MH/s 64.96 MH/s 223/0/6 (100.00)
521m [2021-10-24 20:35:39.378] GPU5 3060Ti 424.60 kH/W 64.96 MH/s 64.96 MH/s 202/0/1 (100.00)
521m [2021-10-24 20:35:39.378] 401.09 kH/W 389.79 MH/s 389.77 MH/s 1365/0/24 (100.00)
521m [2021-10-24 20:35:39.378]
521m [2021-10-24 20:35:39.378] COST: 0.98 KWH (0.08USD) POOL HASHRATE: 381.77 MH/s
October 24, 2021, 07:42:51 PM
To install v1.19 on HiveOS, after 1.16, follow these steps
Then type miner restart and hit return
Weird thing with 1.19
Code:
cd /hive/miners/teamblackminer/1.16/
mv TBMiner TBMiner.1.18.bak
wget https://github.com/sp-hash/TeamBlackMiner/releases/download/v1.19/TeamBlackMiner_1_19_Ubuntu_18_04_Cuda_11_4.tar.gz
tar -xf TeamBlackMiner_1_19_Ubuntu_18_04_Cuda_11_4.tar.gz
cp /hive/miners/teamblackminer/1.16/TeamBlackMiner_1_19_Ubuntu_18_04_Cuda_11_4/TBMiner /hive/miners/teamblackminer/1.16/
>It seems that TBMiner is arlready executable (type ls then hit return, TBMiner should be green)
>if not, run this command
chmod +x TBMiner
mv TBMiner TBMiner.1.18.bak
wget https://github.com/sp-hash/TeamBlackMiner/releases/download/v1.19/TeamBlackMiner_1_19_Ubuntu_18_04_Cuda_11_4.tar.gz
tar -xf TeamBlackMiner_1_19_Ubuntu_18_04_Cuda_11_4.tar.gz
cp /hive/miners/teamblackminer/1.16/TeamBlackMiner_1_19_Ubuntu_18_04_Cuda_11_4/TBMiner /hive/miners/teamblackminer/1.16/
>It seems that TBMiner is arlready executable (type ls then hit return, TBMiner should be green)
>if not, run this command
chmod +x TBMiner
Then type miner restart and hit return
Weird thing with 1.19
Code:
21m [2021-10-24 21:02:47.279] POOL HASHRATE: VARDIFF
While the value was at like 2000 when I started the miner, then it switched to VARDIFF October 24, 2021, 05:03:46 PM
Again an answer besides the question.
Are you a robot?
Are you a robot?
You can read readmes?
--cuda-devices [2,3,4] for example if you don´t want to use gpu 0 and 1 or what is your problem?
@sp_
I have problem with mining eth/etc+zil over rustpool.xyz. I can´t connect to the pool.
TBMiner --algo ethash --hostname eu1-zil.rustpool.xyz --port 8008 --wallet %ETH_WALLET% --server_passwd %ZIL_WALLET%@eth.2miners.com:2020 --worker_name %computername%
Other miner work and etc or eth only also. Maybe TBMiner has problem with zil.rustpool.xyz?
October 24, 2021, 04:32:04 PM
Again an answer besides the question.
Are you a robot?
Are you a robot?
October 24, 2021, 04:15:56 PM
Removing the stales is easy, you just have to run with a low --xintensity.
Like --xintensity 24
v1.16 has a high stale rate on --xintensity -1 . v1.18/v1.19 has halve the stales from 1.16 but hashing at the same speed (--xintensity -1)
Like --xintensity 24
v1.16 has a high stale rate on --xintensity -1 . v1.18/v1.19 has halve the stales from 1.16 but hashing at the same speed (--xintensity -1)
October 24, 2021, 03:53:59 PM
--cl-devices [2,3,4]
I'm under the impression that you don't understand what I'm looking for..
But nevermind; I tested TBM on another rig; hashrate is much higher than GMiner, but it creates too many stales. I know this is by design, and only to be used with pools that pay for stales, but most pools don't, so I'm better off with GMiner anyway..
May check TBM later if things get better..
Jump to: