The more GPUs the more instable my OCs

ProFlo999 · November 2, 2021, 3:04am

I have 8 RX580 cards, run the latest HiveOS, mining on ethermine pool and with teamredminer.

I don’t know how to explain the issue better but it seems to the more GPUs I add to my rig, the more unstable it becomes.

For example, if I mine with GPU 1, 2, 3 the rig is stable, each GPU running at 30MH/s, same if I run GPU 4,5,6. The same of course applied if I run 6,7,8.

But as soon as I connect all of the GPU 2 becomes unstable, meaning as the miner starts this GPU mines at 0MH/s and the rig crashes a few seconds after starting. I lowered all the OCs for all GPUs, even set the GPU back to factory settings but the issue prevailed.

Might it be an issue with the PSU?

Mining rig: BT360 motherboard (above 4G decoding enabled) , 8 RX580, 2 EVGA 1300W Supernova PSU, 8GB of DDR6 RAM, Kingston SSD 64GB

JeanPaulBuloz · November 3, 2021, 4:09am

Have you tried adding one by one and let them run for some hours at any step? Did you disabled CSM in bios mobo ?

ProFlo999 · November 3, 2021, 4:55am

Yes, I did. I added one GPU at a time and let it mine for 20+ minutes each time, Hashrate was stable and no invalid shares at any step. Disabling CSM, I am actually not sure if this has been done already, I will be back home in around 5 hours and will check the BIOS settings. Thanks for the effort to help me in the meantime!

JeanPaulBuloz · November 3, 2021, 4:56am

CSM always gave me troubles… good luck

ProFlo999 · November 3, 2021, 12:58pm

It’s has been disabled all along, so probably also not that’s the cause ;(

JeanPaulBuloz · November 3, 2021, 1:27pm

How many card keep stable together ?..let me research about MB config

ProFlo999 · November 3, 2021, 2:21pm

It remains completely stable until I add 3-4 GPUs. If I mine with 5-8 GPUs the rig keeps on mining but each GPU will be more sensitive to OC settings and will mine 2-7 MH less than it did before. And as soon as I add the 9th GPU one of the GPU completely stop mining with all miner I tried (Ethermine, nano miner, teamredminer). It will basically show 0.00 MH for 1-2 mins until it crashes, which in redlined means “GPU dead”.

I just played around with all mobo settings, I use Gen1 Speed, mining mode is enabled, above 4G coding is enables and CSM is diabled.

Besides I just flashed the problematic GPU with another RX 580 with Hynix memory BIOS that worked fine and now I can still enter the BIOS but the OS won’t boot normally, it just shows a black screen and will reboot 3-4 mins after powering on, but I have a feeling I can fix this issue. I will play around with the settings tomorrow I have a feeling I can fix it by just trying enough.

JeanPaulBuloz · November 3, 2021, 2:26pm

Reflash original bios. I don’t think your problem goes for GPU bios at all. Check your MB bios battery voltaje, should be around 3v.

ProFlo999 · November 3, 2021, 2:44pm

The problem is the rig won’t boot, making it impossible to reflash.

Where can I find the mb voltage?

JeanPaulBuloz · November 3, 2021, 3:10pm

You have to take out the flashed card, To boot up. Those gpu are dual bios ?

ProFlo999 · November 3, 2021, 11:02pm

Nope, single bios

ProFlo999 · November 4, 2021, 2:10pm

Managed to fix it, but the old problem still remains, at least 1 of the 9 GPUs will always show hashrate of 0.00 MH and die soon after booting the rig.

I reset all the PCI slots to Gen 2, then Gen 1, then to auto, but that doesnt seem to be the reason.

Then I tried to downset OCs, but neither that worked, no matter how low I set it, also running the GPU at stock settings didn’t solve it.

I am really running out of options to make any alterations…

JeanPaulBuloz · November 4, 2021, 2:17pm

What miner are you using ? Post first auto tune miner screen to see you OC , post setup miner config from FS screen to see arguments

ProFlo999 · November 4, 2021, 2:30pm

Teamredminer

The miner start immediately after booting, DAG files loading normally and all looks good

![2|666x500]

But as soon as the first hasrates show up one GPU is always at 0MH

And within 1-2 mins after one of the GPUs will sho was dead

Here you can see my current OCs. I tried with higher and lower OCs, but that didn’t help. I am using amdmemtweak – REF 30, but I tried without to exclude this as problemfactor, which didn’t help. (upload://jApRe2kSvXjw3loG03wbiENHgkz.jpeg)