Did I break my card?

[Grea]

As previously shared the F state script is going to set your f state to 1551, which is already set by Hive OS.

Are you setting SOC at 1200? Any SOC voltage? fwiw: I set 850 for most.

You are chasing a GPU, slot, cable, power, overclock, issue…not an OS or Driver issue.

[Obigunshinobi]

so, tried different tweaks to the OC, and system failed to recognized the card until this morning with the original OC settings…happy ending story? no at all

now 20 or so M/s. what is this? that’s it AMD worst gpu for mining as ever, sorry for all the fans out there

Nvidia running for more than 2 years without issues.

the only LHR accident on the way could make you fall out but for the rest, TOP.

[Grea]

If it was my card, it would be pulled from that rig so the other GPUs could keep mining full rate uninterrupted.

Then I would work with the 6800 in a separate test bench/PC. It has run “properly”, so it likely can run properly with the right environment.

Hybrid rigs do add a few more variables. That GPU is presenting enough on its own.

[Alain]

The RX6800 cards are crap to mine with. I had two that I switched from Hiveos to W10, but no difference. One card ended up crashing completely, never to start up again. The other one is hit and miss if it is recognised when starting Windows.

I even switched motherboards from an Asrock B450 to a BTC110. No difference. The three 6700XT’s on the same rig run flawlessly. I will never buy another RX6800.

[sferrariba]

I’ve made a whole bunch of 6800 rigs, all work perfect but the asrock phantom gaming ones.

[Grea]

While I don’t have any ASRock Phantoms, I’ll add, Sapphire 6800s were difficult to get sorted and even now require extra effort.

[Obigunshinobi]

quick update:

card run in windows 10 if…i switch to 2nd Bios mode, and mining on Nbminer without OC settings on with this stable results:

not happy, but the fact that 1 Bios mode work and the other one doesn’t mmmmmmmmmm something went tits up when i tried to download the driver or something, i know it doesn’t make sense but try to hear me out:

• receive the card, plug in it, card not recognized, find the Hiveon guide how to update the driver on hiveos, selected the last driver available, boom, all good, overclocked the card, 61 m/s 110-120 watt, running along another LHR card, Nbminer, not too shabby.
• decided to add 3 more card to the rig, nvidia non LHR, driver problems started, try different things like plug the card alone directly on the motherboard, switch the pcie ports to Gen2, enable 4g on the MOBO, starting to dig into BIOS and stuff like this(no flashing).
• take Grea advice, removed the card from the rig and run it in win10.

things that i have noticed:

• the fan of this card don’t spin until is too late i think, the card turns extremely hot and i think that could be a reason for the crashes so if someone knows how to change this at the root and not through a OC software i would highly appreciate it.
• the bios switch part, it seems that the OC Bios is the problem, how do i bring back as the begin?

thanks all for your feedback, this crying noob miner higly appreciate

[Obigunshinobi]

another noob question:

so i’m running in win10 with no OC for now and these are the stats of the card on the radeon software:

my problem as soon i put OC on the core and memory i have the crash(usually), is it possible that i crash because the system for mining is actually using a lot of core and voltage?
i mean, my goal is to have less power consumption now, so my main focus is to low the usage of the core and the voltage as a consequence of less utilisation of card’s parts. But, is it possible that, by lowering the core, which is used now a stupid speed this will cause the crash?

[Grea]

Great that is it runs in something

If I remember correctly, you should be under clocking the core in Windows and maxing out the memory freq. Not sure if current Radeon drivers make a difference, all my 6800’s are in Hive since the f-state scripts were put out there this summer.

In radeon software try
1100min-1200max Core
2148-2150 Memory
Should be able to back core voltage under 793mV
I’d leave the fans max if there is any suspected cooling.
Save to Radeon Profile at each stage that is stable as you work toward it.

You should be able to load a fresh BIOS in that slot from Hive or Windows with the right settings/tools.

[Obigunshinobi]

ok, after running all morning i came back from lunch and i had the iconic “green screen” of death, everything still running apparentely but obviously i rebooted, windows start and display available.

how would you describe that?

man i just wanna cry

[Grea]

@Obigunshinobi

how would you describe that?

RMA opportunity?

[Obigunshinobi]

really?

what if they found out that i overclocked??

[Grea]

Inside warranty, has been replaced in most instances if you have not cracked open the GPU. In some jurisdictions, those are replaced as well.

Does not hurt to check.

fwiw: My only GPU RMA in years was a recent RTX 3090. Eventually all my 6800s worked in stable fashion.

[Obigunshinobi]

hi all,

quick update, RMA raised and card “repaired” by Gigabyte, well not sure why but the card still no working well when overclocked, BUT, if mining without OC these are the results: 58MH at 129W.

not sure what to think at this point, it may worth says that this is a OC card version so maaaaaaaaaaaybe is already overclocked and doesn’t need additional overclocking
I DON’T KNOW.

[sferrariba]

A card that says its OC version it either means it clocks higher than the same card non-oc, or just higher than reference. What we actually do for mining is the oposite of overclock, we underclock and undervolt cards. The underclock should nlt.make the card crash, but the undervolt might do so. Try one setting at a time, starting with fan speed, then memory speed (that you overclock) then core clock, then if all works well, mingle with the voltages