Ninjastic
PostEdited versionsQuotes to this post
Post
Topic
Board Mining (Altcoins)
Re: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v11.9 (Windows/Linux)
by
Ursul0
on 18/08/2018, 08:57:18 UTC
Quote from: jackbox on August 17, 2018, 04:11:06 PM
Quote from: bitgameSolo on August 17, 2018, 03:19:03 PM
Quote from: Ursul0 on August 17, 2018, 08:11:25 AM
Quote from: AzzAz on August 17, 2018, 12:06:15 AM
Quote from: Ursul0 on August 16, 2018, 12:47:16 PM
@Claymore
It seems there's a correlation between using temps/fan control from the miner and "NVML: cannot get fan speed, error 999 (an internal driver error occurred)" errors, causing application restart (worker restart in my case)
The rig is Win10 with 10*1060*6gb and and latest drivers. I'm using -mclock and -pwlim with no issues running v11.9, but as soon as I set -tt into anything it will lead to the error above within minutes.

I have better results with MSI AB than Claymore on Nvidia

Well... and my conclusion from the experience with MSI AB: do anything you must in order to avoid installing AB on your workers.

Anyways I had serious stability issues with this 10*1060 rig (lately it never held over 5-6 hours of constant work) , it was behaving like it had some problems with electricity/raisers/contacts.
It was constantly restarting/and occasionally hung in a way that only power down/up would make it recover. Reset wasn't doing the trick.

Fresh system image and NO AB install and now I have it running stable at 24.1MH at actually higher mclock than it was before and no issues whatsoever for 21+ hours.

I also have a couple rigs that restart every 4-6 hours. Is there a log file in win7 that would give insight as to whether specifically AB or other driver was at fault? I've been assuming it was just an OC/ temp thing, but hard to figure out which card is at fault so I end up lowering OC on all of them.

Have you looked at the Claymore log files to see what is causing the issues? Claymore log files are very detailed.
speaking of logs- mine were unclear. mostly it looked like power failure - normal operations and then sudden restart/end of log. it was likely a combination of driver, temps control, AB OC profiles issues. also 10 nvidia gpus is a factor, there were various scenarios that would slow down the machine or make logonui 50% (100% of one of the two 1151 celeron cores) or hung on reboot (low level driver issues).
now without AB with plim -35 and mem +750 set in claymore it had 0 restarts past two days.
Note: setting -tt to any value produces driver error something about failure to read temps, but with stock settings its doing fine, the temps are a bit higher than my target of 70C though...



edit: AB is bad, just as it always was. cleaning up the mess it may create could be tough, better never to install it on a worker at all.