GPU power and hashrate halved for no reason

[34gl3]

Hello,

I have a weird issue with my rigs. I’m mining ETC with nVidia GTX 1060 3GB cards. 8 cards in each rig, 4 rigs atm. The issue started around 4-5 days ago when hashrate suddendly droped to 50%. I used to have around 22 mh per GPU, but now only GPUs 0, 6 and 7 are having old hash rate, and GPUs 1-6 are having around 4 or 5 mh. Exactly the same happend on all 4 rigs, and I actually didn’t change anything for more than 2 weaks already…

My OC: proc +150, memory +1500, PL = 70

I have also noticed that GPUs 1-6 are not going over 50 watts… even if I put OC to 0.

I have tried everything from upgrading, rebooting, etc. Anybody had same issue?

Thanks in advance!

Regards

[Grea]

It appears you are showing:

• (2) different Kernel versions
• (3) different HiveOS versions
• (2) different drivers
• (1) miner which does not happen to be on the recommended list. (not saying it does not work, I have used it myself)
• 3rd GPU on “ZOTAC” is not showing fan which points to a potential riser/hardware/overclock issue.

I’d suggest:

• troubleshooting from a baseline of the latest stable kernel release, with latest version of your miner of choice.
• It is interesting that GPU0, GPU6, GPU7 are consistently high.
• GPU1-5 all have lower performance.

Would you happen to be using 1x/4x risers?

[34gl3]

Thank you for your reply.

I cannot use any other miner (eg. T-rex) as from my expirience only nanominer works on 3gb cards (only 3MB memory left tho). T-rex stopped working for me a month ago.

I have just pushed apt-get update && upgrade, selfupgrade, nvidia-driver-update… reboot and its still the same. I have also tried hive-replace --list with latest stable hiveos.

I have all cards on same risers. Look at picture below, this hash drop happend suddendly a few days ago, without me touching anything.

It’s funny that all 4 rigs got the same issue with hash drop on cards 1-6

[Grea]

@34gl3

I have all cards on same risers. Look at picture below, this hash drop happend suddendly a few days ago, without me touching anything.

Sounds very DAG increase related, but I have to say, I know nothing of ETC mining

[34gl3]

Might be. But the funniest part is, if I randomly switch my cards to another slots, there will always be “normal” hash rate at position 0, 6 and 7. And if I remove those 3 positions, I’m left with 5 cards doing 5 mh

I have also thought of hardware issue, but what are the chances of failing 4 rigs on same pci slots at once? I would say close to none

[Grea]

@34gl3

I would say close to none

Agree completely, unless there is a conflict with similar hardware and the new conditions

[34gl3]

Agree. But like I said, I havent touched anything and the drop happend on all 4 rigs at once I don’t ever update/upgrade all of them at once. I am more like, try it on first rig and if it works move further

[Grea]

@34gl3

I am more like, try it on first rig and if it works move further

Same strategy here

[34gl3]

best strategy.
Well thank you for your input! Well appreciated.
I guess I will slowly shut down those 4 rigs and focus on my ETH farm. Fortunately I have 10 rigs full of 6gb-8gb 1070 gtx and p106 cards. Happy mining!

[34gl3]

I forogot to show this pic. Slots from 1-5 are not getting near PL again same on all 4 rigs.