I have some issue now, how to remove the DPM setting? 
- Hive OS web interface
- Select your worker/rig
- Overclocking tab
- going to have to play with edit/duplicate/restore in my experience.
(this is assuming they don’t show up in your “all” dial option for overclocking)
what miner ?
Has anyone here using Biostar Extreme Gaming RX 6800 ? can you share the settings?
I still have this “problem”. Need to resave the OC (changing the fan settings is enough) after a reboot to save ~10watts per card. Did anybody solve this already?
I don’t get why scripts are need to overclock efficently the RX 6800 XT, after some reaserch I get about 62.50 @98w
My 6800 XT sample size is too small, but with the few I have:
- I agree, none my 6800/6800XT’s require scripts to run stable >61MH/s with TRM.
My 6800 sample size is statistically significant.
In my experience, It is not going to hit 100w and you need to measure at the wall to get a real number. The GPU’s under report dramatically:
hey, did you get a solution? I have the same problem with my Asrock RX 6800, everytime the rigs boot, the hashrate drops
Unfortunately, my sample won’t consistently lock in >=62MH/s with miner restarts and adjustments at lower power settings. I’d dial it down and just restart it in a small rig, but since it spends it’s time a larger rig I am not going to restart mining 10 times to get it to perform at the detriment of 5-11 other GPUs 
Good luck with yours!
I’m still on the 0.6-210@211010 (with the F-state script) HiveOS version and haven’t upgraded. Truth is, I’ve actually been quite happy with it. I haven’t rebooted in a month and it is still at 99.47% acceptance rate, with only 1 invalid share for the entire month.
I was hoping for it to crash soon so I could finally upgrade and add another card. But since it’s running smoothly now, I would rather not touch it (my settings are the same as what I posted here last October).
In terms of physical setup, my cards are about 65 cm apart (around 2.5 inches) from shroud to backplate (using a tape measure) while one card is directly plugged on to the board I also use a 2000 watt automatic voltage regulator (output voltage at 220-240v) – it’s the red box next to the monitor.
This here’s an old photo. The rig was still running on Windows back then. It’s basically the same setup now but on HiveOS. Mobo is a Biostar TB250-BTC Pro, 8 GB DDR4-2400, and a Xigmatek X-miner 1800w PSU. Absolutely reliable and durable parts; I should’ve bought more last Feb but I cheaped out. Apart from the RAM, these have all been out of stock for months now. If you have a chance to get them, grab them quick!
2 Likes
What are the miners using? TRM?
Please tell me the numerical values of the SoC frequency and mV settings for reference.
Immediately turn up the fans on that GPU, make sure it is not sitting over the top of a power supply.
You may be able to back down earth of the voltage settings you set by 25. May need to back the Core freq down if you try this, but you can likely up the memory freq a bit.
Many posts I in this thread with various suggestions.
Good luck.
1 Like
i set soc mv 850
Gigabyte cards run hot out of the box. I would suggest replacing the thermal pads.
1 Like
Thanks. I designed a PSU duct to pump the heat out away from the GPU. 3D printing it to put on tonight I hope. I got the temps down to 80c by just putting it into a cold room.
It ran like this for 30 minutes then went back to 438 MHs (GMiner).
I am aware of the occasional bump up to 70 MHs per card, this is definitely not that. Any ideas?