I noticed on my system when I apply overclocks after a reboot, the power draw from the 6800s drops like 5-10w in software and verified at the wall. The hash rate doesn’t change. This drops about 30w at the wall for me typically for my 3 6800s. The overclock change can be as simple as adjusting 1 fan and doesn’t have to be on all cards.
Does anyone else observe this or know whats going on? It would be great if it can do this without me needing to re-applying overclocks.
I got same situatiom as you, when reboot the amd-info show 750mv at 675mv core clock voltage, if applying any overclock simple as you said the fan percentage it will be droping to 668mv at amd-info. This happen after the F state version update, if downgrade to previous version it will be back to normal reboot startup with 668mv amd-info without reapplying any oc.
Which previous version do you use?
I have also experienced this exact issue. Each time the rig is restarted for whatever reason, I change the SoC VDDmax, mV from 800 to 801 or other way around, just to get the 10w drop on the cards
hiveos-0.6-210-beta@210914, but you will lost the F state advantage. My sapphire rx6800 cant seem have other way than manually reapplying overclock with F state advantage version to get lower wattage or just leave it at high wattage, without F state advantage its only manage to get 58mh.
What vbios are you using?
the hiveos wattage reporting for amd cards is way off, so dont have to worry about hiveos, more about at the wall. hiveos shows 410w, but at the wall is 640w. using a corsair ax1600i, g4400, 8gb single stick, 120gb ssd and 4x 6800xt. no additional fans or accessories. sometimes when all the cards are showing 110w usage each, the wall is still at 640w
0.6-211@211029 introduced “Reduced power usage on AMD RDNA2 RX 6x00 series”.
Do you know if it’s just what hiveos shows on screen, or did it also improve at the wall?
Hello Great Guys,
Please find below my OC for 6800, 6800 XT and 1 6900 XT
Almost 7 days running with any issues.
Pool : Ethermine
Miner : NBminer
im not really sure, cos i didnt really check at the wall, only until recently, one of the updates stabilized all the 6800 non-xt, then i started adjusting the SOC to 810mV, saw abit of drop at the wall.
Of course there is less power also “from the wall”! :]
I have experience that you need to multiply reported consumption by approx. x1.35 to get real values. There is also power consumption from riser, that’s probably the difference between reported and “wall” consumption.
Update your Hive Kernel or run the F - state scripts.
If that is your only GPU in the rig, I would update the kernel by researching “hive-replace -y --stable” initiated via Hive Shell.
You are also at the 6800 maximum memory overclock value. That is not a good plan for most GPUs.
Hello everyone,
first of all, many thanks for all the information here in the thread!
I’m using the following settings with gminer/ethermine and my hashrate is with ~62.5 seems quite ok (especially for the asrock). BUT I get always an average load >1 and <2. Any ideas what I could do?
My System:
BTC B250C OEM (5.12 03/26/2021)
4 × Intel® Core™ i5-6400T CPU @ 2.20GHz AES
Model ATA Emtec X150 120GB 120GB
Many thanks in advance!
Are you using a USB dongle to run HiveOS? Are you doing CPU mining as well?
No I’m using a ssd for hive os and I’m not doing CPU ming as well. I’m using a default flightsheet for ethermine/gminer. Also I’m using two 1000watt PSU’s.
I’ve just tweaked the OC a litte bit but my sysload is still high and rising.
Core Clock: 1350
CV: 680
MCV: 800
MC: 1075
MV: 1300
SoC Frequenz: 900
Maybe its a riser problem after all!?
after hours of debugging with: downgrade, fresh install, with one card, with all cards, with no cards, riser replacement etc. I think I found the solution.
It seems that there is a problem with my intel chip and the mainboard bios (ASRock skylake gpe6F). I came up with it because “top” showed me in the terminal, that a kworker process used 60-80% resources. I don’t use asrock but a btc board but it seems to be the same problem.
After applining this to my rig, the load is since some hours under 0.5
https://forums.linuxmint.com/viewtopic.php?t=223180
I will also write a new thread with this. I think this could be very usefull for other users!
Glad you found a solution to your problem but any cpu load issues should not come from GPUs. You can tell by stopping the miner.
I see you have low temps around 58. Mine is always at 70. Are you setting the fan at 75% all the time or is it controlled by autofan ?
