Post
Topic
Board Mining (Altcoins)
Re: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v11.6 (Windows/Linux)
by
xunl
on 03/04/2018, 01:37:37 UTC
Old Env: AMD Driver Version is 17.3 (prior to Blockchain beta), WattMan is installed, Windows 10 Pro 1703. Mix 4x AMD RX580 and 4x nVidia GTX1070 GPUs, Claymore 9.8.
I have used this claymore command line argument to control AMD GPU setting for long time. I found it was great because you can automate the miner and GPU setting during  reboot.
-cclock 1150 –mclock 2100 –cvddc 900 –mvddc 900
But since I upgraded to Blockchain beta (to resolve hashrate decrease/DAG issue), I cannot use Claymore to control AMD GPU setting anymore. I have to manually set it every time in WattMan for each RX580 after reboot. Automation is gone. If I just use Claymore command argument to set it, the setting will reset to default randomly after mining some time. When setting resets to default, you have much longer hashrate but use a lot more power and temperature is much higher.
Current Env: AMD Driver Version is Blockchain beta, WattMan is installed, Windows 10 Pro 1703. Mix 4x AMD RX580 and 4x nVidia GTX1070 GPUs, Claymore 11.2
I have tried some versions of AMD Adrenalin drivers with Claymore 10.x and 11.x, but none of them can maintain setting consistently for long time (10+ hours) w/o reset.
I heard some people use AMD Adrenalin 18.3 drivers with Claymore 11.x achieved -cclock 1150 –mclock 2100 –cvddc 900 –mvddc 900 consistently (4+ days w/o crash). Can you provide me your env settings?
What version of AMD driver you use?
Do you have AMD WattMan installed?
You Claymore version?
Do you mine with RX580?
Another reason I haven’t tried Claymore 11.5 or 11.6 is because someone said claymore will regenerate DAG very hour when they start DEVFEE mining. That would be a serious stability issue. Is that true?
Thanks

FYI:
I have mod/flash some setting into bios so I can automate the startup process. But I found it is very inconvenient because 1) everytime you want change the setting, you have to mod it again. 2) down time is too long if you have lots of GPUs on a rig and many rigs. command line argument (software way vs. hardware way) is far more easier and quicker.